JP5121046B2 - Information processing apparatus and method, program, and information processing system - Google Patents

Description

  The present invention relates to an information processing apparatus and method, a program, and an information processing system, and more particularly, to an information processing apparatus and method, a program, and an information processing system that handle advertisements.

  Some Web servers that provide data of a predetermined Web page (hereinafter simply referred to as a page) to a browser provide data of a page on which an advertisement is arranged.

  Some of these Web servers constitute a content matching system that selects and displays the page content, that is, the advertisement that matches the content, from the advertisements registered by the advertiser.

  In many cases, a contractor other than the advertiser operates the content match system by commissioning from the advertiser.

  For example, the advertisement distribution management network system described in Patent Document 1 manages one or more commercial sites, one or more affiliate sites linked to the commercial sites, and distribution of advertisements to affiliate sites and payment management of advertisement insertion fees. An advertisement distribution management network system in which the advertisement distribution management site to be operated is arranged on the network, the means for extracting the keyword by analyzing the vocabulary and other contents of the contents of the advertisement posting page, and the advertisement ( Text, images, videos, audio, barcodes, and other creatives), and displays advertisements that match the content of the homepage.

  Further, in a mail distribution system that can add advertisement information stored in a server to mail sent to a mail account through a communication line such as the Internet network, there are a plurality of advertisement information added to the mail. Some of the advertisement information is appropriately selected and added based on information registered in advance for each user and / or provision condition information of an advertiser who wants to provide advertisement information (Patent Document 2). reference).

JP-A-2005-242711

JP 2004-158023 A

  When selecting an advertisement that matches the content, it is necessary to extract the feature of the content. When trying to extract a word representing content features from content sentences, words that are difficult to appear in other sentences are extracted as words representing content characteristics.

  On the other hand, if you sell the right to display an ad for a specific query (search keyword), the more frequently used the search is, the more often the ad is displayed, Will be sold at a higher price.

  However, if a word representing a feature is extracted from a sentence, a word that does not appear in other sentences is extracted. Therefore, a query that is used more frequently in a search is less likely to be extracted as a word representing a feature. Tend.

  Furthermore, for a user who searches a page by a specific query and browses the page, it can be said that the query represents the characteristics of the page being browsed.

  In the case of selling the right to display an advertisement matching a content that is an advertisement for a predetermined feature word, in order to sell the right at a higher price, the frequency with which the advertisement is displayed, that is, the feature There is a need to increase the frequency with which words are extracted.

  The present invention has been made in view of such a situation, and enables an advertisement corresponding to a word to be displayed on a page including a word that is a query that is frequently used for a search. It is.

The information processing apparatus according to the first aspect of the present invention multiplies the number of times a query for searching a page is used for a search by the number of search results, which is the number of searched products, Calculating an average value and a standard deviation of products obtained as a result of multiplication for each of the queries, a target value of an average value and a target value of a standard deviation, an average value and a standard deviation of products, and the query By applying a predetermined operation to the product for each, a calculation means for calculating a set value set as a value indicating the likelihood of appearance in each sentence of the query, and a universal appearance for each word And a generation unit that generates a query reflection dictionary by merging the set of queries in which the set values are calculated with a standard dictionary including a value indicating ease.

  In this way, it is possible to extract a query as a feature word, so that an advertisement corresponding to the word is displayed on a page including a word that is a query that is frequently used for a search. Will be able to.

  The generation means includes merging the set of queries in which the setting values are calculated into the standard dictionary including a value that is an occurrence cost and that indicates a universal appearance probability for each word. Thus, the query reflection dictionary can be generated.

  The generation means can generate the query reflection dictionary by rewriting the occurrence cost of a word in the standard dictionary to the set value of the query that is the same as the word.

  The calculation means includes the product C of the number of times used in the search and the number of products searched, the average value AVE of the product C, the standard deviation SD of the product C, the target value NAVE of the average value, and the target value of variance The set value of each of the queries can be calculated by applying the calculation represented by the formula (product C−average value AVE) / standard deviation SD / 3 × target value NSD + target value NAVE to NSD.

  Extraction means for extracting feature words from the text on the page using the query reflection dictionary can be further provided.

  The extraction means can preferentially extract the query as the feature word.

  By doing in this way, more advertisements according to the query can be displayed.

  Acquisition means for acquiring an advertisement for a product or service having the same characteristics as the characteristics indicated by the characteristic words extracted from the text on the page can be further provided.

  Therefore, an advertisement corresponding to the word can be displayed on a page including a word that is a query that is frequently used for search.

  The acquisition means can acquire a product or service advertisement having the same characteristics as the characteristics indicated by the feature words by searching for the feature words.

In the information processing method according to the first aspect of the present invention, the calculation means multiplies the number of times a query for searching a page is used for the search and the number of search results, which is the number of products searched. And calculating the average value and standard deviation of the products obtained as a result of multiplication for each of the plurality of queries, and determining the target value of the average value and the target value of the standard deviation, and the average value and standard deviation of the product When the by applying a predetermined operation on the product of the query respectively, calculates a set value which is set as a value indicating the appearance and ease in the query each sentence, the generator, the words for each of the The method includes a step of generating a query reflection dictionary by merging the set of queries in which the setting values are calculated with a standard dictionary including values indicating the likelihood of universal appearance.

  A query can be extracted as a feature word, and therefore, an advertisement corresponding to the word can be displayed on a page including a word that is a query that is frequently used for a search.

The program according to the first aspect of the present invention multiplies a computer by the number of times a query for searching a page is used for the search and the number of search results, which is the number of products searched, Calculating the average value and standard deviation of the products obtained as a result of multiplication for each of the queries, and determining the target value and standard deviation target value of the predetermined average value, the average value and standard deviation of the product, By applying a predetermined operation to the product for each query, a setting value set as a value indicating the likelihood of appearance in each sentence of the query is calculated, and the universal appearance probability for each word By merging the set of queries for which the set values have been calculated with a standard dictionary including a value indicating, a process including a step of generating a query reflection dictionary is performed.

  A query can be extracted as a feature word, and therefore, an advertisement corresponding to the word can be displayed on a page including a word that is a query that is frequently used for a search.

In the information processing apparatus according to the second aspect of the present invention, the number of times a query for searching a page is used for the search is multiplied by the number of results of the search, which is the number of products searched for, An average value and a standard deviation of products obtained as a result of multiplication for each of the queries are calculated, a predetermined average target value and a standard deviation target value, a product average value and a standard deviation, and the query By applying a predetermined operation to the product for each, the set of queries for which a set value set as a value indicating the likelihood of appearance in each sentence of the query is calculated, for each word Extraction means to extract feature words from page sentences and extraction from page sentences using query reflection dictionary generated by merging into standard dictionary including values indicating universal appearance probability It comprises obtaining means for obtaining an advertisement for goods or services having similar characteristics to the characteristics indicated by the characteristic word that was.

  An advertisement corresponding to a word can be displayed on a page including a word that is a query that is frequently used for a search.

  The extraction means can preferentially extract the query as the feature word.

  The acquisition means can acquire a product or service advertisement having the same characteristics as the characteristics indicated by the feature words by searching for the feature words.

In the information processing method according to the second aspect of the present invention, the number of times a query for searching a page is used for the search by the extraction means is multiplied by the number of search results, which is the number of products searched. A product average value and standard deviation obtained as a result of multiplication for each of the plurality of queries are calculated, and a predetermined average target value and standard deviation target value, and a product average value and standard deviation are calculated. And by applying a predetermined operation to the product for each of the queries, the set of queries for which a set value set as a value indicating the likelihood of appearance in each sentence of the query is calculated. using the query reflected dictionary generated by merging the standard dictionary containing a value indicating the universal occurrence ease of each word, extracts feature words from the text of the page, by the acquisition means, Comprises obtaining the advertisement of goods or services has similar features and characteristics represented by the characteristic word extracted from the text of the over-di.

  An advertisement corresponding to a word can be displayed on a page including a word that is a query that is frequently used for a search.

A program according to the second aspect of the present invention is obtained by multiplying a computer by the number of times a query for searching a page is used for a search and the number of search results, which is the number of products searched. The average value and standard deviation of the products obtained as a result of multiplication for each of the queries are calculated, the target value of the average value and the target value of the standard deviation, the average value and standard deviation of the product, By applying a predetermined calculation to the product for each query, a set of queries for which a set value set as a value indicating the likelihood of appearance in each sentence of each query is calculated, for each word Using query reflection dictionaries generated by merging into a standard dictionary that includes values indicating the ease of universal appearance, feature words are extracted from page text and extracted from page text. To perform processing comprising the step of acquiring the advertisement of goods or services has similar features and characteristics represented by the characteristic word is.

  An advertisement corresponding to a word can be displayed on a page including a word that is a query that is frequently used for a search.

The information processing system according to the third aspect of the present invention multiplies the number of times a query for searching a page is used for a search by the number of search results, which is the number of searched products, Calculating an average value and a standard deviation of products obtained as a result of multiplication for each of the queries, a target value of an average value and a target value of a standard deviation, an average value and a standard deviation of products, and the query By applying a predetermined operation to the product for each, a calculation means for calculating a set value set as a value indicating the likelihood of appearance in each sentence of the query, and a universal appearance for each word A generation unit that generates a query reflection dictionary by merging the set of queries in which the setting values are calculated into a standard dictionary including a value indicating ease, and a page using the query reflection dictionary Comprising extracting means for extracting a characteristic word from the text, and an acquisition means for acquiring the advertisement of goods or services has similar features and characteristics represented by the characteristic word extracted from the text of the page.

  An advertisement corresponding to a word can be displayed on a page including a word that is a query that is frequently used for a search.

  As described above, according to the first aspect of the present invention, a query can be extracted as a feature word.

  Further, according to the first aspect of the present invention, an advertisement corresponding to a word can be displayed on a page including a word that is a query that is frequently used for a search.

  According to the second aspect of the present invention, an advertisement can be displayed on a page.

  In addition, according to the second aspect of the present invention, an advertisement corresponding to a word can be displayed on a page including a word that is a query that is frequently used for a search.

  Furthermore, according to the 3rd side surface of this invention, the advertisement according to the word can be displayed on the page containing the word which is a query with high frequency used for a search.

  Embodiments of the present invention will be described below. Correspondences between the configuration requirements of the present invention and the embodiments described in the detailed description of the present invention are exemplified as follows. This description is to confirm that the embodiments supporting the present invention are described in the detailed description of the invention. Accordingly, although there are embodiments that are described in the detailed description of the invention but are not described here as embodiments corresponding to the constituent elements of the present invention, It does not mean that the embodiment does not correspond to the configuration requirements. Conversely, even if an embodiment is described here as corresponding to a configuration requirement, that means that the embodiment does not correspond to a configuration requirement other than the configuration requirement. It's not something to do.

The information processing apparatus according to the first aspect of the present invention multiplies the number of times a query for searching a page is used for a search by the number of search results, which is the number of searched products, Calculating an average value and a standard deviation of products obtained as a result of multiplication for each of the queries, a target value of an average value and a target value of a standard deviation, an average value and a standard deviation of products, and the query A calculation means (for example, a generation cost calculation unit in FIG. 6 ) that calculates a set value that is set as a value indicating the likelihood of appearance in each sentence of the query by applying a predetermined calculation to each product. 91) and a generation means for generating a query reflection dictionary by merging the set of queries in which the set values are calculated into a standard dictionary including a value indicating the universal appearance probability for each word (example) If provided with a dictionary merge portion 92) and Fig.

  Extraction means (for example, the analysis unit 94 in FIG. 6) for extracting feature words from the text on the page using the query reflection dictionary can be further provided.

  An acquisition unit (for example, the advertisement creation unit 96 in FIG. 6) that acquires advertisements of products or services having the same characteristics as the characteristics indicated by the characteristic words extracted from the text on the page can be further provided.

The processing performed by executing the information processing method or the program according to the first aspect of the present invention includes the number of times a query for searching a page is used for the search and the result of the search that is the number of products searched. The product is obtained as a result of multiplication for each of the plurality of queries, and an average value and a standard deviation are calculated, and a predetermined average target value and a standard deviation target value are calculated. By applying a predetermined calculation to the value and the standard deviation and the product for each of the queries, a setting value set as a value indicating the likelihood of appearance in each sentence of the query is calculated (for example, FIG. 10 step S14), by merging the set of queries for which the set values have been calculated into a standard dictionary including a value indicating the universal appearance probability for each word, Generating a (e.g., step S15 in FIG. 10) includes the step.

In the information processing apparatus according to the second aspect of the present invention, the number of times a query for searching a page is used for the search is multiplied by the number of results of the search, which is the number of products searched for, An average value and a standard deviation of products obtained as a result of multiplication for each of the queries are calculated, a predetermined average target value and a standard deviation target value, a product average value and a standard deviation, and the query By applying a predetermined operation to the product for each, the set of queries for which a set value set as a value indicating the likelihood of appearance in each sentence of the query is calculated, for each word Extraction means (for example, the analysis unit 9 in FIG. 6) extracts a feature word from a sentence on a page using a query reflection dictionary generated by merging with a standard dictionary including a value indicating a universal appearance probability. Comprising a), and an acquisition means for acquiring the advertisement of goods or services having similar characteristics and features represented by the feature words extracted from the text of the page (e.g., ad creation unit 96 of FIG. 6).

The processing performed by executing the information processing method or the program according to the second aspect of the present invention includes the number of times a query for searching for a page is used for the search and the result of the search that is the number of searched products. The product is averaged and the standard deviation obtained as a result of multiplication for each of the plurality of queries is calculated, and the target value of the average value and the target value of the standard deviation and the product average are calculated. By applying a predetermined calculation to the value and the standard deviation and the product for each of the queries, a setting value that is set as a value indicating the likelihood of appearance in each sentence of the query is calculated. Using query reflection dictionaries generated by merging a set of queries into a standard dictionary containing values that indicate the likelihood of universal appearance for each word, feature words are extracted from the text on the page. The advertisement (for example, step S32 in FIG. 11), and the advertisement of the product or service having the same characteristics as the characteristics indicated by the characteristic words extracted from the text on the page (for example, step S34 in FIG. 11) is obtained. Including.

The information processing system according to the third aspect of the present invention multiplies the number of times a query for searching a page is used for a search by the number of search results, which is the number of searched products, Calculating an average value and a standard deviation of products obtained as a result of multiplication for each of the queries, a target value of an average value and a target value of a standard deviation, an average value and a standard deviation of products, and the query A calculation means (for example, a generation cost calculation unit in FIG. 6 ) that calculates a set value that is set as a value indicating the likelihood of appearance in each sentence of the query by applying a predetermined calculation to each product. 91) and a generation unit that generates a query reflection dictionary by merging the set of queries in which the set values are calculated with a standard dictionary including a value indicating the universal appearance probability for each word. For example, using the dictionary merging unit 92 in FIG. 6 and the query reflection dictionary, the extraction means (for example, the analyzing unit 94 in FIG. 6) for extracting feature words from the page text and the page text. It has an acquisition means (for example, advertisement creation part 96 of Drawing 6) which acquires the advertisement of goods or services which have the same feature as the feature shown by the above-mentioned feature word.

  FIG. 1 is a diagram showing a configuration of an information processing system according to an embodiment of the present invention.

  The market provider provides the store with a service that provides the user with a page for selling the product by entrusting from the store, which is a corporation or an individual (natural person) who sells the product.

  The market provider includes a Web server 11, an affiliate server 12, a product database (DB) 13, and an advertisement database (DB) 14.

  The Web server 11 executes, for example, a so-called Web server program such as Apache HTTP Server or Internet Information Services manufactured by Microsoft Corporation in the United States, and based on a procedure defined in HTTP (Hypertext Transfer Protocol), such as text or image Provides documents such as HTML (Hypertext Markup Language) or XML (Extensible Markup Language) that include various objects. For example, a document is data of a web page.

  Hereinafter, a web page is simply referred to as a page.

  The Web server 11 provides the user with data of a page commissioned by a store and a page for accepting an application for offering a product or service (service). By entrusting from a large number of stores, the Web server 11 provides the user with a virtual market consisting of a page for selling a large number of products and a page for accepting applications for the provision of a large number of services.

  The affiliate server 12 provides a so-called affiliate service. That is, the affiliate server 12 is an object linked to a page for selling a product or a page for accepting an application for provision of service, and an object for accessing the page is arranged on the user's page, and this page is browsed. When another user purchases a product on a page that sells a product via this link or makes an application for provision of a service, a service is provided in which a reward is paid to a user who has placed an object on his / her page.

  Hereinafter, a user who places an object on his / her page and obtains a reward is also referred to as an affiliate.

  The affiliate server 12 issues a script for providing an affiliate service. The affiliate server 12 provides the issued script to the affiliate device. Execution of this script provides an advertisement that matches the content of the page on which the product or service advertisement is placed.

  The advertisement that matches the content of the page on which the advertisement for the product or service is placed is an advertisement for the product or service having characteristics similar to the characteristics of the page on which the advertisement for the product or service is placed. Here, similar features refer to the same features, similar features, or approximate features.

  For example, an advertisement for a product or service includes an object for accessing a page for selling the product or a page for accepting an application for provision of service.

  Hereinafter, an advertisement that matches the content of the page on which the advertisement for the product or service is arranged is referred to as a content match advertisement.

  For example, a script for providing a content match advertisement affiliate service is Javascript (Java is a registered trademark).

  The affiliate server 12 is a device of a user who browses a page on which an advertisement is provided, which is provided by the affiliate, and receives an advertisement in response to a request from a device that executes a script placed on the page. provide.

  Hereinafter, a user who browses a page provided by an affiliate is also referred to as a page browsing user.

  Hereinafter, a page on which an object for accessing a page for selling a product or a page for accepting an application for providing a service using an affiliate service is referred to as an affiliate page.

  Further, the affiliate server 12 generates a dictionary necessary for extracting a feature word that is a word indicating the feature from the sentence. The affiliate server 12 uses the generated dictionary to extract a feature word that is a word indicating the feature from a sentence on an affiliate page on which an advertisement for a product or service is placed.

  FIG. 2 is a diagram for explaining a dictionary generated in the affiliate server 12.

  A set A on the left side of FIG. 2 is a set of queries (search keywords) for searching pages. For example, the query belonging to the set A is a query (search keyword) for searching for a page that a market provider provides to a user on consignment of a store, and that sells a product or receives a service application. ).

  The right to display advertisements for individual queries belonging to this set A is sold to corporations or individuals. For example, the right to display an advertisement for a query is sold to a store.

  The set B on the left side in FIG. 2 that intersects with the set A is a set of feature words that are words indicating the features of the text extracted from the text on the page using the standard dictionary. Here, the standard dictionary is a dictionary that is created based on a corpus of correct Japanese sentence examples and machine-learned. Note that the standard dictionary includes an occurrence cost indicating the ease of universal appearance for each word.

  The intersection of the set A and the set B, that is, the product set C of the set A and the set B, is a word that is a query for searching the page, and is a feature word extracted from the text of the page by the standard dictionary. It consists of a word.

  That is, since words belonging to the product set C are extracted as feature words using the standard dictionary, when a sentence includes this word, an advertisement corresponding to the word is displayed on the page of the sentence.

  In principle, a word belonging to the set A but not belonging to the product set C is not extracted as a feature word using the standard dictionary, and even if the word is included in the text of the page, the advertisement corresponding to the word is not extracted. Do not show.

  Therefore, as shown on the right side of FIG. 2, a feature word that is a query is extracted from the text on the page using a set of standard dictionaries and queries (search keywords). That is, the number of words belonging to the product set C is increased. In other words, the number of words belonging to the set A and not belonging to the product set C is decreased.

  By doing in this way, more words among the words used as a query (search keyword) for searching a page are extracted as feature words. Therefore, an advertisement corresponding to the word can be more reliably displayed on the page including the word that is one of the queries used for the search.

  As a result, most of the queries used for the search can be sold as a word with a high frequency of displaying the advertisement, and the right to display the advertisement for the word can be sold.

  Returning to FIG. 1, the product database 13 stores information on products or services. For example, the merchandise database 13 stores information related to merchandise sold on a page provided to a user and a service for accepting an application for provision on the page in response to a request from a store.

  For example, the product database 13 receives a command from a member who is a store via the Web server 11 for registration, editing, addition, etc. of information related to a product sold by the member or a service to be provided. Store new information or update stored information.

  FIG. 3 is a diagram illustrating an example of product data (records) stored in the product database 13. In the product database 13, a product name that is data indicating the name of a product or a service, a store ID that is data specifying a store that sells the product or a store that provides the service, and a product or service are specified. The product ID that is the data to be stored, the product price that is the data indicating the price of the product or the charge (consideration) of the service, and the product description that is the text data describing the product or the service are stored.

  In addition, the product database 13 includes a product page address that is data indicating the address of a page for selling a product or a page for accepting an application for provision of a service, a product image indicating the presence or absence of an image of a product or an image that symbolizes a service. A product image page address which is data indicating an address of a page on which a product or service image is posted when there is a no flag and an image symbolizing the product image or service is stored.

  Further, the product database 13 stores a sellable flag indicating whether or not the product is in stock at the store (whether or not the product can be sold), or whether or not the service can be provided immediately.

  The product database 13 also includes a consumption tax flag indicating whether it is a so-called internal tax or an external tax, a postage flag indicating whether or not a postage is necessary, and a credit card indicating whether or not payment is possible using a credit card. An available flag is stored.

  Furthermore, the product database 13 includes a shop of the year flag indicating whether or not a store that sells products or a store that provides services has been recognized as an excellent store, the price of the product or the fee for services when using an affiliate. Affiliate usage rate, which is the data showing the interest rate of remuneration, sales start time, data showing the time to start accepting applications for sales of products or services, and termination of acceptance of applications for sales of products or services Data indicating the time to sell, the end time of sales, the number of reviews indicating the number of evaluations of goods or services by critics or users who have purchased products or received services, and data indicating the average of the evaluations The average of reviews is stored.

  In this way, if the product or service identified by the same product ID is handled at multiple stores identified by different store IDs, the product or service handled at different stores is considered to be different. , Data of each product or service is managed.

  The product database 13 stores a product name, a product ID, a product price, a product description, and the like of each of a plurality of products handled in the store for one store ID that identifies one store. May be.

  The advertisement database 14 stores advertisement data. That is, the advertisement database 14 stores advertisement data to be displayed on a page provided by a market provider in accordance with a contract with a store or a corporation or an individual other than the store.

  FIG. 4 is a diagram illustrating an example of advertisement data (record) stored in the advertisement database 14. For example, the advertisement database 14 includes a title that is text data of the headline of the advertisement, an explanatory text that is text data of text explaining the product or service of the advertisement, an address that is data indicating an address of the page of the product or service of the advertisement, In addition, a keyword or the like which is text data indicating characteristics of the advertisement (product or service) is stored.

  More specifically, for example, the title indicates the name of a product sold by a registered member or a service provided by a registered member, or a catch phrase of the product or service.

  When the affiliate server 12 searches the advertisement database 14 using a feature word extracted from the sentence as a query, the keyword stored in the advertisement database 14 is the same as the feature word extracted from the sentence in the affiliate server 12; Or when they are similar, that is, when the keyword is similar to the feature word, the advertisement of the keyword is retrieved from the advertisement database 14.

  Further, for example, the advertisement database 14 is an object for accessing data for displaying an image of the product or an image symbolizing the service, or a page of a person (including a member) who sells the product or provides the service. Stores the data to be displayed.

  Returning to FIG. 1, the Web server 11, the affiliate server 12, the product database 13, and the advertisement database 14 are connected to each other by a LAN (Local Area Network) 15.

  A router 16 is further connected to the LAN 15, and the Web server 11 and the affiliate server 12 are connected to the network 17 via the LAN 15 and the router 16.

  The network 17 is a network including a wired or wireless general-purpose line or a dedicated line, and is, for example, the Internet, a LAN, or a WAN (Wide Area Network).

  In principle, direct access to the product database 13 and the advertisement database 14 via the network 17 is prohibited.

  More specifically, the Web server 11 and the affiliate server 12 are connected to the LAN 15 so as to respond to access via the network 17, and the product database 13 and the advertisement database 14 are directly accessed via the network 17 by a firewall. It is connected to the LAN 15 so that it cannot be done. However, the Web server 11, the affiliate server 12, the product database 13, and the advertisement database 14 are connected to the LAN 15 so that they can directly access each other.

  Further, the product database 13 or the advertisement database 14 may be constructed in either the Web server 11 or the affiliate server 12.

  Stores that are corporations or individuals that sell products or provide services are each provided with a terminal device 18-1 or a terminal device 18-2. Each of the terminal device 18-1 and the terminal device 18-2 is, for example, a personal computer, and through a network 17, through authentication using a registered member ID and password, The service page provided from the store is updated, or the product or service data stored in the product database 13 is updated.

  Each user operates using the user terminal device 19-1 or the user terminal device 19-2.

  Each of the user terminal device 19-1 and the user terminal device 19-2 is, for example, a stationary or so-called notebook personal computer, a mobile phone, or a dedicated terminal device, and the user server 20-1 of another user or The user server 20-2 or the web server 11 is accessed, the page data provided from the user server 20-1, the user server 20-2, or the web server 11 is received, and the page is displayed.

  For example, the user operates the user terminal device 19-1 to transmit a member ID and password for specifying a member to the Web server 11, thereby allowing each of a page for selling each of a number of products and a number of services. Log in to a virtual market consisting of a page that accepts an application for provision, and purchase a product or apply for a service from the page of a store in the market.

  The user server 20-1 or the user server 20-2 executes, for example, a so-called Web server program such as Apache HTTP Server or Internet Information Services manufactured by Microsoft Corporation, based on a procedure defined in HTTP, etc. Documents such as HTML or XML that include various objects are provided.

  The user server 23-1 or the user server 23-2 may be, for example, a server of an Internet connection service provider, a server that provides a blog service, or a so-called rental server.

  In addition, a provider (hereinafter referred to as an affiliate service provider) different from the market provider provides the affiliate service alone or provides the affiliate service on behalf of the market provider. Also good. In this case, the affiliate service provider includes the affiliate server 12. When an affiliate service provider acts as an affiliate service provider by a market provider, the affiliate service provider and the market provider are connected by a dedicated line or connected by a VPN (Virtual Private Network) or the like. The

  Hereinafter, when it is not necessary to individually distinguish the user terminal device 19-1 or the user terminal device 19-2, they are simply referred to as the user terminal device 19. Further, when there is no need to individually distinguish the user server 20-1 or the user server 20-2, they are simply referred to as the user server 20.

  FIG. 5 is a block diagram illustrating an example of a hardware configuration of the affiliate server 12.

  In the affiliate server 12, a CPU (Central Processing Unit) 51, a ROM (Read Only Memory) 52, and a RAM (Random Access Memory) 53 are connected to each other by a bus 54.

  An input / output interface 55 is further connected to the bus 54. The input / output interface 55 includes an input unit 56 composed of a keyboard, mouse, microphone, etc., an output unit 57 composed of a display, a speaker, etc., a storage unit 58 composed of a hard disk and a non-volatile memory, and a communication unit 59 composed of a network interface. A drive 60 for driving a removable medium 61 such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory is connected.

  In the affiliate server 12 configured as described above, the CPU 51 loads the program stored in the storage unit 58 into the RAM 53 via the input / output interface 55 and the bus 54 and executes the program, for example. A series of processes are performed.

  The program executed by the affiliate server 12 (CPU 51) is, for example, a magnetic disk (including a flexible disk), an optical disk (CD-ROM (Compact Disc-Read Only Memory), DVD (Digital Versatile Disc), etc.), a magneto-optical disk, Alternatively, it is recorded on a removable medium 61 that is a package medium made of a semiconductor memory or the like, or provided via a wired or wireless transmission medium such as a network 17 such as a local area network, the Internet, or digital satellite broadcasting.

  The program can be installed in the affiliate server 12 by installing the removable medium 61 in the drive 60 and storing it in the storage unit 58 via the input / output interface 55. The program can be installed in the affiliate server 12 by being received by the communication unit 59 via a wired or wireless transmission medium and stored in the storage unit 58. In addition, the program can be installed in advance in the affiliate server 12 by storing the program in the ROM 52 or the storage unit 58 in advance.

  Since the hardware configuration of the Web server 11, the user terminal device 19, and the user server 20 is the same as the hardware configuration of the affiliate server 12, description thereof is omitted.

  Next, with reference to FIG. 6, the structure of the function implement | achieved by the affiliate server 12 which executes a program is demonstrated.

  When the affiliate server 12 executes the program, the merge dictionary generation unit 81, the query set storage unit 82, the standard dictionary storage unit 83, the merge dictionary storage unit 84, the advertisement association unit 85, the association advertisement storage unit 86, and the script issue The unit 87, the advertisement providing unit 88, and the reward processing unit 89 are realized.

  The merge dictionary generation unit 81 generates a merge dictionary by merging a set of queries with a standard dictionary including a value indicating the universal appearance probability for each word.

  That is, the merge dictionary generation unit 81 calculates a setting value set as a value indicating the likelihood of appearance in the sentence of the query from the frequency at which the query for searching the page is used for the search. A merge dictionary, which is an example of a query reflection dictionary, is generated by merging a set of queries for which setting values have been calculated into a standard dictionary that includes a value indicating a universal appearance probability.

  The query set storage unit 82 manages a predetermined area of the storage area of the storage unit 58 and stores a set of queries in that area. More specifically, the query set storage unit 82 searches the storage unit 58 for a page provided by a market provider by consignment of a store, and a page for selling a product or receiving a service application. The query set used in the above is stored, and the stored query set is read in response to a request from the merge dictionary generation unit 81.

  For example, the Web server 11 executes a search CGI (Common Gateway Interface) application and stores a set of queries as a search log. The query set stored in the storage unit 58 by the query set storage unit 82 is obtained from the Web server 11 stored as a search log.

  FIG. 7 is a diagram illustrating an example of a set of queries stored in the storage unit 58 by the query set storage unit 82. A set of queries includes, for each of a plurality of queries, a query that is a search keyword (word), a search count that indicates the number of times the query was used for the search, a product page or service searched by the search by the query The number of hit products indicating the number of pages in the URL, etc. are included.

  In this case, each of the queries is preferably a word that is a morpheme. For example, the number of searches indicates the number of searches in a predetermined unit period (for example, one week or one month). Further, for example, the number of searches may be the total number of searches so far, without defining a period.

  For example, the number of hit products may be set to a value that is the total of the same products, or may be set to a value that is the total number of hit pages.

  The standard dictionary storage unit 83 manages a predetermined area of the storage area of the storage unit 58, and stores a standard dictionary including a value indicating a universal appearance probability for each word that is a morpheme in the area. Yes.

  FIG. 8 is a diagram illustrating an example of a standard dictionary stored in the storage unit 58 by the standard dictionary storage unit 83. For each word, the standard dictionary includes the word, the occurrence cost, which is a value that indicates the likelihood of the word, the part of speech of the word, the reading of the word, and the pronunciation of the word. Yes.

  The merge dictionary storage unit 84 manages a predetermined area of the storage area of the storage unit 58 and stores the merge dictionary generated by the merge dictionary generation unit 81 in the area.

  FIG. 9 is a diagram illustrating an example of the merge dictionary stored in the storage unit 58 by the merge dictionary storage unit 84. In the merge dictionary, for each of a plurality of words, a word, an occurrence cost that is a value indicating the easiness of appearance of the word, a part of speech of the word, a reading of the word, and a pronunciation of the word, and the word The price indicating the selling price of the right to display the advertisement is included.

  The occurrence cost of a word in the merge dictionary, which is a query belonging to a set of queries, is different from the occurrence cost of that word in the standard dictionary.

  Returning to FIG. 6 again, the advertisement association unit 85 associates the page on which the advertisement is to be placed with the advertisement of the product or service. In this case, for example, the advertisement association unit 85 associates an advertisement of a product or service having the same characteristics as the characteristics of the page with the address of the page. More specifically, for example, the advertisement association unit 85 associates an advertisement of a product or service having the same characteristics as the characteristics of the affiliate page with the address of the affiliate page.

  The association advertisement storage unit 86 manages a predetermined area in the storage area of the storage unit 58, and stores the advertisement of the product or service associated with the page in the advertisement association unit 85 in the area. For example, the association advertisement storage unit 86 stores the advertisement of the product or service associated with the affiliate page in the area of the storage unit 58.

  The script issuing unit 87 issues a script for providing a content match advertisement affiliate service. The issued script is an affiliate who is a user who uses the affiliate service, and the user terminal device is connected to the user server 20 of the affiliate who wants to place an advertisement including an object for accessing the page for selling the product on the affiliate page. 19 is provided.

  The script provided to the user server 20 is executed by the accessing user terminal device 19 when the page on which the script is arranged is accessed, and an advertisement is requested from the affiliate server 12 by executing the script.

  The advertisement providing unit 88 provides an advertisement in response to an advertisement request from the user server 20 by executing the script.

  The reward processing unit 89 provides a service on a page where a page browsing user who browses an affiliate page purchases a product on a page for selling a product or accepts an application for providing a service via a link on the affiliate page. When applying, a process of giving a reward to the affiliate who provides the affiliate page is executed.

  The merge dictionary generation unit 81 includes a generation cost calculation unit 91 and a dictionary merge unit 92.

  The generation cost calculation unit 91 calculates a setting value that is set as a value indicating the likelihood of appearance of a query in a sentence from the frequency with which a query for searching a page is used for the search. That is, the generation cost calculation unit 91 generates, for each query belonging to the set of queries stored in the storage unit 58 by the query set storage unit 82, a set value based on the frequency with which the query was used for the search. Calculate the cost.

  The dictionary merging unit 92 generates a merge dictionary by merging a set of queries for which a generation cost, which is a set value, is calculated, into a standard dictionary including a value indicating the universal appearance probability for each word.

  The advertisement association unit 85 includes a sentence acquisition unit 93, an analysis unit 94, a feature word determination unit 95, and an advertisement creation unit 96.

  The sentence acquisition unit 93 acquires a sentence of a page to which an advertisement is to be associated. For example, the text acquisition unit 93 acquires from the user server 20 the HTML source (document described in the HTML format) of the address of the page to which the advertisement is to be associated, and the main text as the main text from the acquired HTML source. The sentence of a predetermined page is acquired by extracting the sentence.

  The analysis unit 94 analyzes the acquired sentence using the merge dictionary and decomposes the sentence into words. For example, the analysis unit 94 decomposes a sentence into words by morphological analysis using a merge dictionary.

  The feature word determination unit 95 determines a feature word that characterizes the sentence from among the decomposed words. For example, the feature word determination unit 95 sets a predetermined number of words as feature words in descending order of occurrence cost from among the decomposed words.

  For example, the feature word determination unit 95 refers to the merge dictionary and prioritizes the query as a feature word from among the decomposed words. Specifically, for example, the feature word determination unit 95 refers to the merge dictionary, and multiplies the occurrence cost for the query among the decomposed words by a predetermined constant to increase the occurrence cost of the query. Thus, the query is preferentially used as the feature word by setting a predetermined number of words as the feature word in descending order of occurrence cost.

  The advertisement creation unit 96 creates an advertisement that matches the feature word. For example, the advertisement creation unit 96 searches the product database 13 or the advertisement database 14 using the feature word as a query, and sets the searched product or service information as an advertisement associated with a sentence (page). That is, the advertisement creating unit 96 associates an advertisement of a product or service having the same characteristics as the characteristics of the text with the text (page).

  The advertisement includes an image of the product, text indicating the name of the product, an object that is a button for accessing a page for selling the product, and the like. That is, the advertisement created by the advertisement creating unit 96 is data.

  If the feature word is a query that sells the right to display an advertisement, the advertisement creating unit 96 associates the advertisement of the corporation or the individual who purchased the right with the sentence (page). By doing in this way, the advertisement of the corporation or the individual who purchased the right to display the advertisement for the predetermined query is surely associated with the document (page) from which the feature word that is the query is extracted.

  In addition, as a result of searching the advertisement database 14, if an advertisement for a product or service having the same characteristics as the characteristics of the text is obtained, the advertisement creating unit 96 uses the obtained advertisement as it is as the text (page). Associate.

  Furthermore, the merge dictionary generation unit 81, the query set storage unit 82, the standard dictionary storage unit 83, and the merge dictionary storage unit 84 may be realized in an information processing device different from the affiliate server 12. In this case, the information processing device different from the affiliate server 12 generates a merge dictionary, and the affiliate server 12 uses the merge dictionary generated by the information processing device different from the affiliate server 12 to use the feature word from the sentence. To create an advertisement that matches the feature word. That is, an information processing system is configured by an information processing apparatus and the affiliate server 12 different from the affiliate server 12.

  Next, merge dictionary generation processing will be described. FIG. 10 is a flowchart illustrating an example of merge dictionary generation processing.

  In step S11, the generation cost calculation unit 91 of the merge dictionary generation unit 81 causes the query set storage unit 82 to read the stored query set. In step S11, the generation cost calculation unit 91 calculates, for each query, a product C of the number of searches A and the number B of hit products included in the set of queries.

  That is, for each query, the generation cost calculation unit 91 hits the number of searches A indicating the number of times the query was used for the search, and the number of product pages or service pages searched by the search by the query. Multiply by the number of products B to calculate the product C of the number of searches A and the number B of hit products.

  In step S12, the generation cost calculation unit 91 calculates the average value AVE and standard deviation SD of the product C calculated in step S11. A plurality of queries belong to the set of queries, and in step S11, the product C is calculated for each of the queries. Therefore, the average value AVE and the standard deviation SD of the plurality of products C are calculated.

  In step S13, the generation cost calculation unit 91 determines the target value NAVE of the average value of occurrence costs to be determined and the target value NSD of variance. For example, in step S13, the generation cost calculation unit 91 sets the target value NAVE of the average value and the target value of the variance so that the average cost and the variance of the word occurrence cost in the standard dictionary have a predetermined ratio. Determine NSD.

  More specifically, for example, in step S13, the generation cost calculation unit 91 causes the average value target value NAVE and the range of variance to be approximately equal to the average value and variance of word occurrence costs in the standard dictionary. The target value NSD is determined.

  Further, for example, in step S13, the generation cost calculation unit 91 has a predetermined magnification such as 0.7 times or 0.8 times with respect to the average value and variance of word occurrence costs in the standard dictionary. The average target value NAVE and the variance target value NSD are determined.

  In the standard dictionary, for each morpheme, an occurrence cost, which is a value indicating the ease of appearance of each morpheme, is defined in advance, and further, the ease of connection of two morphemes with respect to a sequence of two morphemes. The connected cost shown is predefined. In order to eliminate ambiguity when performing morphological analysis, the morphological analysis engine that performs morphological analysis often uses an algorithm called the “minimum cost method” that selects the solution that minimizes the sum of the concatenation cost and the generation cost. Used.

  Therefore, when it is desired to extract a query preferentially from the result of morphological analysis, it is preferable to set the cost of occurrence of the query low in advance. For example, the target value NAVE of the average value of the occurrence cost of the query and the target value NSD of the variance are the average value of the word occurrence cost of the standard dictionary, a predetermined magnification such as 0.6 times or 0.7 times the variance. It is decided to become.

  Further, the generation cost calculation unit 91 may determine an average value target value NAVE and a variance target value NSD as predetermined values. For example, the generation cost calculation unit 91 determines an average target value NAVE to a predetermined value of 2000, and determines a variance target value NSD to a predetermined value of 1000.

  In step S14, the generation cost calculation unit 91 determines the occurrence cost of each query from the product C, average value AVE, standard deviation SD, target value NAVE of average value, and target value NSD of variance. That is, the generation cost calculation unit 91 calculates an occurrence cost, which is an example of a setting value set as a value indicating the likelihood of appearance in a sentence of the query, based on the frequency at which the query for searching the page is used for the search. decide.

  For example, in step S14, the generation cost calculation unit 91 calculates the product C, the average value AVE of the product C, the standard deviation SD of the product C, the target value NAVE of the average value, and the target value NSD of the variance by the equation (1). The occurrence cost of each query is calculated by applying the indicated operation.

Occurrence cost = (product C-average value AVE) / standard deviation SD / 3 x target value NSD + target value NAVE
... Formula (1)

  In step S15, the dictionary merging unit 92 of the merge dictionary generating unit 81 merges the set of queries for which the occurrence cost has been determined with the standard dictionary. That is, in step S15, the dictionary merging unit 92 merges the set of queries for which the occurrence costs have been determined and the standard dictionary, generates a merge dictionary, and stores the generated merge dictionary in the merge dictionary storage unit 84. Then, the merge dictionary generation process ends.

  In other words, in step S15, the dictionary merging unit 92 merges a set of queries for which occurrence costs, which are examples of setting values, are calculated into a standard dictionary that includes a value indicating a universal appearance probability for each word. Thus, a merge dictionary that is an example of a query reflection dictionary is generated.

  For example, in step S15, the dictionary merging unit 92 merges a set of queries in which the occurrence cost of each query is calculated into a standard dictionary including the occurrence cost indicating the ease of universal appearance for each word. Generate a merge dictionary. More specifically, for example, the dictionary merging unit 92 generates a merge dictionary by rewriting the occurrence cost of a word in the standard dictionary to a setting value of the same query as that word.

  As described above, a setting value (for example, occurrence cost) set as a value indicating the likelihood of appearance in the sentence of the query is calculated from the frequency at which the query for searching the page is used for the search. A merge dictionary, which is an example of a query reflection dictionary, is generated by merging a set of queries into a standard dictionary that includes a value (for example, an occurrence cost) indicating a universal appearance probability for each word.

  Next, with reference to the flowchart of FIG. 11, description will be given of a process for associating a sentence with an advertisement, which is executed with reference to the merge dictionary generated as described above. In step S31, the text acquisition unit 93 of the advertisement association unit 85 acquires the target text.

  For example, in step S <b> 31, first, the sentence acquisition unit 93 acquires from the user server 20 the HTML source of the page to which the advertisement is to be associated. More specifically, for example, the sentence acquisition unit 93 acquires from the user server 20 an HTML source of an affiliate page on which an advertisement is to be placed by an affiliate page display process described later.

  For example, the text acquisition unit 93 acquires, from the user server 20, an HTML source of a page that is provided by the Web server 11 and that is to be associated with an advertisement.

  In step S31, the sentence acquisition unit 93 acquires a sentence on a predetermined page by extracting a sentence such as a main body, which is a main sentence, from the acquired HTML source.

  In step S32, the analysis unit 94 of the advertisement association unit 85 applies the morphological analysis process using the merge dictionary stored in the merge dictionary storage unit 84 to the sentence acquired in step S31. A sentence to which morphological analysis processing using a merge dictionary is applied is broken down into words.

  In morphological analysis, a set of rules for Japanese grammar, the target language, and a dictionary, which is a word list with information such as parts of speech, are used, and sentences written in natural language (in this case, sentences ) Are divided into columns of morphemes, ie, the smallest units (words) that have meaning in the language, and the part of speech of each morpheme is determined.

  In the morphological analysis in step S32, a merge dictionary generated by combining a set of queries and a standard dictionary that is created based on a corpus of correct Japanese sentence collections and machine-learned is used.

  In the morphological analysis, the words of the sentence are segmented, the endings of the used words are processed, and the part of speech of the words is identified. The correct answer is the one that is grammatically correct from the part-of-speech order obtained as a result.

  A method using rules, a method using a probabilistic language model, or the like is employed in the division of words in a sentence, that is, the determination of word boundaries. Specifically, for example, a statistical method based on a minimum cost method or a Hidden Markov Model (HMM) can be adopted.

  Morphological analysis is one of the basic technologies of natural language processing using computers such as computers. In morphological analysis, knowledge of target language grammar (gathering of grammar rules) and dictionary (word list with information such as parts of speech) are used as information sources, and sentences written in natural language are used as morphemes (Morpheme, roughly In other words, it is divided into columns of the smallest unit having meaning in the language, and the part of speech of each morpheme is determined.

  Generally speaking, a part of speech to which a word cut out from a sentence belongs is examined using a dictionary, and a grammatically correct part is taken as a correct answer from the part of speech part obtained as a result. In general, many natural languages have restrictions on the connection of parts of speech (for example, in Japanese, a case particle cannot come after a verb. “Do not open the door”, etc.). By utilizing this property, the boundary between words is discriminated. Specifically, there are two methods that use this property: a method using rules, and a method using a probabilistic language model.

  FIG. 12 is a diagram illustrating an example of an analysis target sentence that is a sentence to be subjected to morphological analysis and a result of the morphological analysis.

  For example, “This year's“ Father's Day Special Feature 2007 ”theme is“ Health and Negirai ”in response to purchasing trends among men over 50 in the“ Easy Market ”. When the morphological analysis process is applied to the sentence to be analyzed, the sentence is “part of speech that is a noun that can be adverbed” “this year”, the part of speech that has been determined to be a participle of linking, “ "", Which was determined to be part-of-speech as a symbol of openness, "Father", which was determined to be part-of-speech as a noun for general nouns, "", which was determined to be part-of-speech as an adjunct to adjunct Determined as “noun” as part of speech that is a noun, “special feature” as part of speech that is a noun of a strange connection, “2007” as part of speech as an unknown word, and as part of speech that is a bracketed symbol “” ”, Etc., are divided into morpheme strings, and the part of speech of each morpheme is determined.

  For example, when morphological analysis is applied to the body of a so-called blog page to extract feature words of the page, as shown in FIG. 13, “Today is Mother's Day. If the chef offers champagne from the Champagne region ... the flowers purchased at the Yoyogi Uehara florist ... "and the morphological analysis processing is applied to the text, the cost of occurrence is 3069. “Today” is determined to be a part of speech that is a noun that can be adverbed, is a noun of a general noun with an occurrence cost of 2944, and is a “part of speech that is determined to be a part of speech that is a particle of a particle” "Mother" determined to be part of speech, occurrence cost is 712, determined to be part of speech that is a coalesced particle, and determined to be part of speech that is a non-independent adverb noun with occurrence cost of 2150 The “day” The cost of occurrence is 1297, and the part of speech is determined to be part-of-speech that is a symbol of a punctuation mark. It is divided into morpheme strings, and the cost of occurrence of each morpheme is obtained. Determined.

  As shown on the right side of FIG. 13, the occurrence cost defined in advance in the merge dictionary indicates the ease of appearance of a certain word. A word with a high occurrence cost is generally a word that rarely appears. Accordingly, it can be said that a word with a high occurrence cost is a feature word that characterizes a sentence.

  Among the morphemes extracted from the example shown in FIG. 13, morphemes that are nouns with high occurrence costs, that is, feature words are “champagne”, “chef”, “champagne”, as shown in the lower right of FIG. , “Yoyogi Uehara”, and “Flower”.

  Returning to FIG. 11, in step S <b> 33, the feature word determination unit 95 of the advertisement association unit 85 determines a predetermined number of nouns or unknown words as feature words in descending order of occurrence cost. That is, in step S33, the feature word determination unit 95 determines a final feature word to be used in the subsequent process from the feature words output by the morphological analysis process in step S32.

  In step S34, the advertisement creation unit 96 of the advertisement association unit 85 searches for the advertisement from the product database 13 or the advertisement database 14 using the feature word as a query.

  For example, when the words “champagne”, “chef”, “champagne”, “Yoyogi Uehara”, and “florist” are determined as feature words in step S33, as shown in FIG. The advertisement creation unit 96 searches for advertisements from the product database 13 or the advertisement database 14 using “champagne”, “chef”, “champagne”, “Yoyogi Uehara”, and “florist” as queries.

That is, for example, as shown in FIG. 15, the advertisement creating unit 96 uses the feature words extracted from the text (termX ₁ , termX ₂ , termX ₃ , termX ₄ , termX ₅ , termX _6. ) As a search term (query), a product or service is searched from the product database 13 or the advertisement database 14, and an advertisement of the product or service obtained as a result of the search is used as an advertisement that matches a feature word extracted from the sentence. .

For example, as illustrated in FIG. 16, the advertisement creating unit 96 causes the analysis unit 94 and the feature word determining unit 95 to extract feature words from the advertisement and extract the feature words from the sentences (in FIG. 16). termX ₁ , termX ₂ , termX ₃ , termX ₄ , termX ₅ , termX ₆ , etc.) (termY ₁ , termY ₂ , termY ₃ , termY ₄ , termY ₅ , termY _{6 in} FIG. 16) The advertisement from which ()) is extracted may be an advertisement that matches the feature word extracted from the sentence.

  In other words, in this case, the advertisement creating unit 96 obtains the probability that the feature word extracted from the sentence and the feature word extracted from the advertisement overlap as the similarity between the sentence and the product or service. The advertisement creation unit 96 sets an advertisement for a product or service having a degree of similarity greater than a predetermined threshold value that matches a feature word extracted from the sentence.

  Returning to FIG. 11, in step S <b> 35, the advertisement creating unit 96 outputs the advertisement obtained as a result of the search to the association advertisement storage unit 86 as an advertisement that matches the sentence, and performs a process for associating the sentence with the advertisement. Ends. The association advertisement storage unit 86 causes the storage unit 58 to store an advertisement that is output from the advertisement creation unit 96 and matches the sentence. That is, the association advertisement storage unit 86 causes the storage unit 58 to store advertisements in association with pages.

  In this way, the text and the advertisement are associated with each other. That is, the advertisement is associated with the page.

  When the page is displayed, the advertisement associated with the page is read, and the advertisement is arranged and displayed on the page. The displayed advertisement is for a product or service with characteristics similar to the characteristics of the page.

  As described above, the page on which the advertisement is placed and displayed is a page for selling a product or a page for accepting an application for providing a service, an affiliate page, or a market provider commissioned by a market provider. It can be a page provided to a user or a store.

  Since the page browsing user of the user terminal device 19 browses a page describing that he / she is interested, the product or service having the same characteristics as the feature is the page browsing user of the user terminal device 19. Likely to attract attention. Therefore, a product or service advertisement having characteristics similar to the characteristics of the page can be viewed by a page browsing user, and a page for selling the product or receiving a service application can be accessed from the advertisement. Increases nature.

  If an advertisement is placed and the displayed page is an affiliate page, an affiliate who provides the affiliate page can place an advertisement having a higher possibility of obtaining a reward on his / her affiliate page.

  From the store, it is possible to access your own product or service on a page that is likely to access the page that sells the product or the page that accepts the application for service provision (the page that is likely to lead to the sale or service of the product). Can be displayed.

  Next, processing for associating a sentence with an advertisement when a word as a query is given priority as a feature word will be described with reference to a flowchart of FIG.

  Since step S51 and step S52 are the same as step S31 and step S32 of FIG. 11, respectively, description thereof is omitted.

  In step S53, the feature word determination unit 95 of the advertisement association unit 85 prioritizes the words included in the query set over the words included only in the standard dictionary, in order of increasing occurrence costs. The number of nouns or unknown words is determined as a feature word.

  That is, a word that is used more frequently in search and for which more products or service pages are searched is preferentially used as a feature word. In other words, words that are more valuable to market providers (words that can earn more profit) are preferentially feature words.

  For example, in step S53, the feature word determination unit 95 refers to the merge dictionary and increases the occurrence cost of the query by multiplying the occurrence cost for the query among the decomposed words by a predetermined constant. Thus, the query is preferentially used as the feature word by setting a predetermined number of words as the feature word in descending order of occurrence cost.

  In this case, for example, in the merge dictionary, the information merged from the query set is associated with a word merged from the query set.

  Further, for example, in step S53, the feature word determination unit 95 determines that the words included in the set of queries are included only in the standard dictionary so that a predetermined number of queries are included in the feature words. Is also given priority and nouns or unknown words are determined as feature words.

  Further, for example, in step S53, the feature word determination unit 95 determines that the query number so that the number of words included in the query set and the number of words included only in the standard dictionary are a predetermined ratio. The noun or the unknown word is determined as the feature word by giving priority to the word included in the set of words over the word included only in the standard dictionary.

  For example, in step S53, the feature word determination unit 95 may determine only a noun or an unknown word that is a query as a feature word without using a noun or an unknown word that is not a query as a feature word.

  The feature word determination unit 95 is included in the query set by referring to the set of queries stored in the query set storage unit 82 and the standard dictionary stored in the standard dictionary storage unit 83. A predetermined number of nouns or unknown words may be determined as feature words by prioritizing words over words included only in the standard dictionary.

  In this way, since words included in a set of queries are prioritized as feature words over words included only in the standard dictionary, it is possible to display more advertisements according to the query. become.

  In the merge dictionary generation process, it is recommended to set the query generation cost low in advance. However, when the sentence is associated with the advertisement based on the feature word, the extracted noun is used when determining the feature word. Alternatively, unknown words are used as feature words in descending order of occurrence costs, and at first glance, it seems that queries are difficult to be extracted as feature words. However, this problem can be avoided by using queries as feature words preferentially.

  Since step S54 and step S55 are the same as step S34 and step S35 of FIG. 11, respectively, description thereof is omitted.

  As described above, words that are included in a set of queries that are used more frequently in searches and that have more products or services as search results are included only in the standard dictionary. Because it is determined to be a feature word in preference to the current word, the product or service advertisement is displayed in the display of the advertisement on the page compared to the case where the feature word is simply set in descending order of occurrence cost. It becomes easy. That is, it is possible to prevent the product or service from being determined as a feature word that cannot be searched, in other words, an advertisement cannot be searched. Will be displayed.

  For example, a product frame, which is an area for displaying an advertisement associated with the page, and an advertisement frame for displaying a pre-signed advertisement are provided on the page, and the product frame has the same characteristics as the features of the page When the service advertisement is displayed, the advertisement can be displayed so that the product frame is filled. When a product line is filled with advertisements, it will increase access to the product or service page from that page, increase the opportunity for the user to view the advertisement for the inventory, and the right to display the advertisement for the inventory. It becomes easy to sell. That is, both the product space and the advertising space are easily filled with the advertisement.

  In addition, advertisements for products or services with similar characteristics to the page's characteristics are displayed, so that the characteristics of the page are emphasized, and an external search engine searches the page with a query similar to the characteristic word. If that happens, the page will be easier to hit. Therefore, the possibility that the page is browsed is further increased. The same applies when searching only the pages provided by the market provider.

  Next, a case where an advertisement associated with the affiliate page is displayed on an affiliate page as an example of a page on which the advertisement is displayed will be described.

  First, with reference to the flowchart of FIG. 18, processing for issuing Javascript, which is a script for providing an affiliate service for content match advertising, by the user terminal device 19 of the affiliate and the affiliate server 12 will be described.

  In step S3001, the user terminal device 19 transmits the member ID and password input from the affiliate through the network 17 (using a browser) and requests login. In response to this request, the affiliate terminal In step S2001, the server 12 authenticates the other party (affiliator) with the member ID and password transmitted via the network 17, the router 16, and the LAN 15, and executes login processing.

  Through this process, the affiliate server 12 authenticates that the affiliate of the user terminal device 19 is a valid member.

  The member ID and password are registered in advance and stored in the affiliate server 12.

  In step S <b> 3002, the user terminal device 19 requests Javascript for the content match advertisement via the network 17.

  In step S2002, the script issuing unit 87 of the affiliate server 12 receives a request from the user terminal device 19. That is, for example, the script issuing unit 87 causes the communication unit 59 that is a network interface to receive a request command transmitted from the user terminal device 19 via the network 17, the router 16, and the LAN 15.

  In step S2003, the script issuing unit 87 of the affiliate server 12 issues a Javascript corresponding to the member ID acquired during the login process. For example, the script issuing unit 87 issues a new Javascript by embedding a value corresponding to a member ID or the like in a Javascript template stored in advance.

  The member ID embedded in Javascript is used as an affiliate ID.

  In step S2003, the issued Javascript is an example of a script for providing an advertisement that matches the content of the page on which the advertisement of the product is placed.

  FIG. 19 is a diagram illustrating an example of Javascript issued in step S2003.

"<Script type =" text / javascript "in Javascript shown in FIG.
language = "javascript">"indicates that this script is Javascript.

  “Width = 720 height = 250” in the Javascript shown in FIG. 19 indicates that the width of the area in which the content match advertisement is displayed is 720 pixels, and the height of the area is 250 pixels.

"Src = http: //xxxx.api.rakuraku.co.jp/xxxxxx? Afid = xxxx = 720 & url ="
+ document.URL + "></iframe><br>");"" http://xxxx.api.rakuraku.co.jp/xxxxxx? "is the affiliate where the content match ad is placed Indicates the URL address of the page, “afid = xxxx” is the affiliate ID that identifies the affiliate, and “+ document.URL” is a process that dynamically acquires the URL address of the affiliate page. Is instructed to execute.

  Returning to FIG. 18, in step S2004, the script issuing unit 87 of the affiliate server 12 sends the issued Javascript to the user terminal device 19 via the LAN 15, the router 16, and the network 17 to the communication unit 59 that is a network interface. Send it.

  In step S3003, the user terminal device 19 receives the Javascript transmitted from the affiliate server 12. In step S3004, the user terminal device 19 describes the Javascript received in step S3003 as a tag on the affiliate page provided by the affiliate, and the Javascript issuing process ends. That is, the user terminal device 19 transmits affiliate page data describing Javascript to the user server 20 by a method such as FTP (File Transfer Protocol), and the user server 20 transmits the Javascript transmitted from the user terminal device 19. Store the data of the affiliate page that describes. Thereafter, the user server 20 provides affiliate page data describing Javascript in response to the request.

  In this way, the affiliate server 12 issues a script for providing an advertisement that matches the content of the page on which the advertisement for the product or service is arranged in response to a request from the user terminal device 19. The issued script is described on a predetermined page provided by the user server 20.

  Note that the content of the page on which the advertisement is displayed may be examined before issuing the script, and the script may be issued when the examination passes. In this case, the address of the page on which the advertisement is displayed is transmitted from the user terminal device 19 (or the user server 20) to the affiliate server 12 together with the script request. The affiliate server 12 stores information indicating whether the examination has passed or not, corresponding to the page address.

  In this embodiment, the user terminal device 19 acquires the script. However, the corporate user server 20 that provides the blog service acquires the corporate ID for identifying the corporate entity, and the blog template. The present invention can also be applied to the case where affiliate information (Javascript) is embedded in a file. In that case, the processing described as being performed by the user terminal device 19 shown in FIG. 18 is executed in the user server 20.

  Next, an affiliate page display process in which Javascript is described in the process described with reference to FIG. 18 will be described. By this process, an affiliate page on which advertisements of products or services having the same characteristics as the characteristics of the affiliate page are arranged is displayed.

  FIG. 20 is a flowchart for explaining the affiliate page display process.

  In step S3101, the user terminal device 19 requests affiliate page data. That is, in step S 3101, the user terminal device 19 transmits an affiliate page access request to the user server 20 via the network 17.

  In step S <b> 1101, the user server 20 receives a request for affiliate page data from the user terminal device 19. In other words, in step S1101, the user server 20 receives an affiliate page access request transmitted from the user terminal device 19 via the network 17, and the requested affiliate page data in response to the access request. Is read.

  In step S <b> 1102, the user server 20 transmits the requested affiliate page data to the user terminal device 19 via the network 17.

  In step S 3102, the user terminal device 19 receives the requested affiliate page data transmitted from the user server 20 via the network 17.

  In step S3103, the user terminal device 19 executes the Javascript described in the affiliate page, and the affiliate server 12 embeds the address of the affiliate page and the affiliate ID (embedded in the Javascript) by executing the Javascript being executed. And request advertisements. In other words, in step S3103, the user terminal device 19 adds an affiliate page address and an affiliate ID, and requests an advertisement from the affiliate server 12.

  That is, the user terminal device 19 transmits the affiliate page address and the affiliate ID to the affiliate server 12 via the network 17 by executing Javascript, and also to the affiliate server 12 via the network 17 in the same manner as the characteristics of the affiliate page. Require feature advertising.

  In step S <b> 2101, the advertisement providing unit 88 receives a request from the user terminal device 19. That is, in more detail, the advertisement providing unit 88 sends the affiliate page address and the affiliate ID transmitted from the user terminal device 19 to the communication unit 59, which is a network interface, via the network 17, the router 16, and the LAN 15. And an advertisement request.

  In step S2102, the advertisement providing unit 88 acquires an advertisement corresponding to the address attached to the advertisement request from the association advertisement storage unit 86.

  In step S2102, the advertisement providing unit 88 embeds the affiliate ID received in step S2101 in the link included in the acquired advertisement.

  In step S2103, the advertisement providing unit 88 sends to the user terminal device 19 that has requested the affiliate server data from the user server 20 via the LAN 15, the router 16, and the network 17 to the communication unit 59 that is a network interface. Send an ad.

  In step S <b> 3104, the user terminal device 19 receives an advertisement transmitted from the affiliate server 12 via the network 17. In step S3105, the user terminal device 19 embeds the advertisement received in step S3104 in the affiliate page received in step S3102.

  In step S3106, the user terminal device 19 displays the affiliate page on which the advertisement is arranged based on the affiliate page data in which the advertisement is embedded, and the affiliate page display process ends.

  As shown in the lower side of FIG. 19, the user terminal device 19 embeds an advertisement requested by the execution of the script and transmitted from the affiliate server 12 in the affiliate page supplied from the user server 20. Then, display the entire affiliate page where the advertisement is placed.

  Thus, the user terminal device 19 displays an affiliate page on which advertisements for products having the same characteristics as the characteristics are arranged.

  Since the page browsing user of the user terminal device 19 browses an affiliate page describing that he / she is interested, a product or service having the same characteristics as the feature is viewed on the page of the user terminal device 19. It is likely to attract the user's attention. Therefore, an advertisement for a product or service having characteristics similar to the characteristics of the page placed on the affiliate page is seen by the page browsing user, and an application for providing a page or service for selling the product is received from the advertisement. The possibility of accessing the page to be accepted increases.

  The affiliate server 12 receives a request for advertisement data by an API (Application Programming Interface) in which input parameters and output parameters are defined, and outputs the requested advertisement data. In this way, the affiliate server 12 can be used more easily.

  For sites with development capabilities that can use the market provider's WebAPI (Application Programming Interface), the information obtained using the market provider's WebAPI is uniquely optimized, as shown on the left side of FIG. Can be displayed.

  On the other hand, as shown on the right side of FIG. 21, by using the content match distribution API that performs the above-described processing, a site or mobile site that wants to use a simple API can use the content match distribution API. By using WebAPI, it is possible to execute the same process as when using WebAPI more easily than when using WebAPI directly.

  A site that wants to use a simple API or a mobile site would prefer a content match distribution API that can be easily used. In particular, it is more effective on mobile sites where capacity and processing capacity are limited.

  If the content of the page on which the advertisement is displayed is examined before issuing the script, and the affiliate server 12 stores information indicating whether or not the examination has passed corresponding to the address of the page, After S2101 and before step S2102, it is determined whether the affiliate page has been screened based on the address of the affiliate page and information indicating whether the screen has been screened, and the affiliate page has been screened. If it is determined that the process has been performed, the processes in steps S2102 and S2103 are performed. If it is determined that the affiliate page has not been examined, the processes in steps S2102 and S2103 are not performed.

  Next, referring to the flowcharts of FIG. 22 and FIG. 23, the product purchase process executed when the affiliate page on which the advertisement of the product having the same feature is arranged is displayed. explain.

  In step S3301, the user terminal device 19 determines whether or not the advertisement is clicked while browsing the affiliate page on which the advertisement is arranged. If it is determined that the advertisement is not clicked, the procedure proceeds to step S3301. Returning to, the determination process is repeated until the advertisement is clicked.

  If it is determined in step S3301 that the advertisement is clicked while browsing the affiliate page on which the advertisement is placed, the procedure proceeds to step S3302. In step S3302, the user terminal device 19 requests the affiliate server 12 for data of a page for selling the product for the advertisement via the network 17.

  In step S <b> 2301, the reward processing unit 89 of the affiliate server 12 receives a request from the user terminal device 19. More specifically, the reward processing unit 89 causes the communication unit 59, which is a network interface, to receive a request from the user terminal device 19 via the network 17, the router 16, and the LAN 15, and accepts the request.

  In step S2302, the reward processing unit 89 includes a link click history including information (affiliate ID) indicating the affiliate who is the user providing the affiliate page where the advertisement is arranged, and information indicating the affiliate. Is stored in a cookie.

  Specifically, the affiliate ID of the affiliate who is the user providing the affiliate page is stored in association with the address (the address of the page described in the URL method) when the request is received in step S2301. Therefore, the reward processing unit 89 analyzes the address to obtain an affiliate ID that is information indicating the affiliate and includes it in the link click history.

  In step S3302, the user terminal device 19 transmits the affiliate ID embedded in the advertisement together with the page data request, and the reward processing unit 89 receives the affiliate ID in step S2301, and in step S2302 The link click history including the received affiliate ID may be stored in a cookie.

  In step S2303, the reward processing unit 89 transmits the cookie storing the link click history to the user terminal device 19. That is, the reward processing unit 89 causes the communication unit 59 that is a network interface to transmit a cookie storing the link click history to the user terminal device 19 via the LAN 15, the router 16, and the network 17.

  In step S3303, the user terminal device 19 receives the cookie storing the link click history transmitted from the affiliate server 12 via the network 17. In step S3304, the user terminal device 19 stores the received cookie in an internal storage unit or the like.

  In step S2304, the reward processing unit 89 of the affiliate server 12 causes the user terminal device 19 to transmit data to be redirected to the advertisement destination address. Note that the address of the advertisement destination corresponds to the address of the page requested in the process of step S3302.

  In step S 3305, the user terminal device 19 receives the redirect data transmitted from the affiliate server 12 via the network 17. In step S <b> 3306, the user terminal device 19 refers to the redirect data and requests the Web server 11 for the data of the advertisement destination page, that is, the page selling the advertisement product, via the network 17.

  In step S4301, the Web server 11 receives a request for data of the advertisement destination page from the user terminal device 19, and transmits data of the advertisement destination page to the user terminal device 19 in step S4302.

  In step S <b> 3307, the user terminal device 19 receives the data of the advertisement destination page transmitted from the Web server 11, that is, the page selling the advertisement product. In step S3308, the user terminal device 19 displays the advertisement destination page.

  In step S <b> 3309, the user terminal device 19 determines whether or not a button for instructing storage of a product to be purchased in a stage before purchase, that is, a so-called shopping cart button is clicked. If it is determined in step S3309 that the shopping cart button has not been clicked, the procedure returns to step S3309 and repeats the determination process.

  If it is determined in step S3309 that the shopping cart button has been clicked, the procedure proceeds to step S3310, and the user terminal device 19 puts the cookie including information (member ID) specifying the page browsing user into the shopping cart. And the information specifying the product for which the is selected is transmitted to the Web server 11 via the network 17. This cookie is transmitted from the Web server 11 in advance when the page browsing user logs in to the site (for example, before executing the affiliate page display process), and is managed by the user terminal device 19. Stored in the stored storage area.

  In step S4303, the Web server 11 receives the cookie including information specifying the user and the information specifying the product selected to be put in the shopping cart, which are transmitted from the user terminal device 19 via the network 17. To do.

  If login or page browsing user registration processing has not been executed, only the information for specifying the product selected to be put in the shopping cart in step S3310 is sent from the user terminal device 19 via the network 17 to the Web. The information is transmitted to the server 11 and the information is received in step S4303.

  In step S4304, the Web server 11 updates the cookie so that the information specifying the selected product is stored in the cookie by, for example, a CGI application. For example, in step S4304, the Web server 11 causes the CGI application to store, in a cookie, information specifying the selected product that has been transmitted as a parameter or the like by the POST method. Alternatively, for example, the Web server 11 stores information specifying the selected product in a cookie by a so-called server-side script.

  In step S 4305, the Web server 11 transmits the updated cookie storing information for specifying the selected product to the user terminal device 19 via the network 17.

  In step S <b> 3311, the user terminal device 19 receives the updated cookie transmitted from the Web server 11 via the network 17. In step S3312, the user terminal device 19 stores the received cookie therein (for example, a storage area managed by a storage unit provided therein).

  Note that, by repeating the processing of step S3309 to step S3312 and step S4303 to step S4305, information specifying each of a plurality of products desired to be purchased is stored in the cookie.

  In step S3313, the user terminal device 19 determines whether or not a so-called purchase button for instructing purchase has been clicked. If it is determined in step S3313 that the purchase button has not been clicked, the procedure returns to step S3313 and repeats the determination process.

  If it is determined in step S3313 that the purchase button has been clicked, the procedure proceeds to step S3314, and the user terminal device 19 includes information specifying the selected product and information (member ID) specifying the user. The cookie is transmitted to the Web server 11 via the network 17.

  In step S4306, the Web server 11 receives a cookie including information specifying the selected product and information specifying the user (member ID) transmitted from the user terminal device 19 via the network 17.

  In step S4307, the Web server 11 confirms the cookie, and executes a purchase process for the product placed in the shopping cart.

  As a result of the purchase process, the store providing the page has accepted the application for a contract for purchasing (buying and selling) the product of the page browsing user. Thereafter, the store that provides the page delivers the product to the page browsing user, and the page browsing user pays the price of the product to the store that provides the page.

  In step S <b> 3315, the user terminal device 19 transmits the cookie storing the link click history and the purchased product information to the affiliate server 12 via the network 17.

  In step S2305, the affiliate server 12 causes the communication unit 59, which is a network interface, to receive the cookie transmitted from the user terminal device 19 via the network 17, the router 16, and the LAN 15.

  In step S2306, the reward processing unit 89 determines whether or not the cookie has a link click history. If it is determined that the cookie has a link click history, the procedure proceeds to step S2307. In step S2307, the reward processing unit 89 gives reward points to the page on which the advertisement is arranged, that is, the affiliate who provides the affiliate page. The affiliate providing the affiliate page is specified by the affiliate ID indicating the affiliate providing the affiliate page where the advertisement is arranged, stored in the link click history.

  In addition, the reward point given in step S2307 is an undetermined one temporarily given, and is decided after the page browsing user who purchased the product pays for the product. The process of determining reward points is executed once a day, for example.

  Furthermore, the confirmed reward points are collected every month, and given to the affiliate as shopping points that can be actually used by the affiliate who provides the affiliate page in that month.

  In step S2308, the reward processing unit 89 updates the cookie so as to delete the link click history from the cookie. In step S 2309, the reward processing unit 89 transmits the updated cookie to the user terminal device 19 via the LAN 15, the router 16, and the network 17.

  In step S 3316, the user terminal device 19 receives the updated cookie transmitted from the affiliate server 12 via the network 17. In step S3317, the user terminal device 19 stores the received cookie, and the product purchase process ends.

  If it is determined in step S2306 that there is no link click history in the cookie, the merchandise purchase processing is terminated without giving reward points to the affiliate or updating the cookie.

  The expiration date of the link click history included in the cookie is 30 days. When the expiration date has passed, the user terminal device 19 substantially deletes the link click history (for example, invalidates the cookie). ). If it is within the expiration date, the link click history is not erased unless the product purchase process is executed.

  In addition, the affiliate who provides the affiliate page via the affiliate page advertisement, even if the page purchase user purchases a product on a page different from the page linked to the affiliate page advertisement. Reward points may be given to

  Furthermore, in the example of FIG. 22 and FIG. 23, the processing in the case of being executed by a personal computer has been described. In this case, since a cookie is not used in mobile, a similar process may be executed using a session ID or the like.

  As described above, when a product is purchased via an advertisement on an affiliate page, reward points are given to the affiliate who is the user providing the affiliate page.

  In this way, when the morphological analysis process is applied to the sentence, the query can be extracted as a feature word. In addition, a setting value that is set as a value indicating the likelihood of occurrence of the query in the sentence of the query is calculated from the frequency at which the query for searching the page is used for the search, and the universal appearance probability for each word When a query reflection dictionary is generated by merging a set of queries for which setting values have been calculated into a standard dictionary that includes values indicating, a word that is a frequently used query is searched An advertisement corresponding to the word can be displayed on the included page.

  A set of queries for which a set value that is set as a value indicating the likelihood of appearance in the sentence of the query is calculated from the frequency at which the query for searching the page is used for the search, is universal for each word. Using a query reflection dictionary generated by merging into a standard dictionary that includes values indicating the likelihood of appearance, feature words are extracted from the page text, and the features indicated by the feature words extracted from the page text When an advertisement for a product or service having similar characteristics is acquired, an advertisement corresponding to the word can be displayed on a page including a word that is a query that is frequently used for a search.

  In addition, a setting value set as a value indicating the likelihood of occurrence of the query in the sentence of the query is calculated from the frequency with which the query for searching the page was used for the search, and the universal appearance probability for each word is calculated. A query reflection dictionary is generated by merging a set of queries for which setting values have been calculated into a standard dictionary including values indicating When an advertisement for a product or service having the same characteristics as the characteristics indicated by the characteristic words extracted from the sentence is acquired, the page containing the words that are frequently used in the search is included in the page. An advertisement corresponding to the word can be displayed.

  The series of processes described above can be executed by hardware or can be executed by software. When a series of processing is executed by software, a program constituting the software executes various functions by installing a computer incorporated in dedicated hardware or various programs. For example, it is installed from a program recording medium in a general-purpose personal computer or the like.

  The program executed by the computer may be a program that is processed in time series in the order described in this specification, or in parallel or at a necessary timing such as when a call is made. It may be a program for processing.

  The embodiments of the present invention are not limited to the above-described embodiments, and various modifications can be made without departing from the scope of the present invention.

It is a figure which shows the structure of the information processing system of one embodiment of this invention. It is a figure explaining the dictionary produced | generated in the affiliate server. It is a figure which shows the example of the data of the goods stored in the goods database. It is a figure which shows the example of the data of the advertisement stored in the advertisement database. 2 is a block diagram illustrating an example of a hardware configuration of an affiliate server 12. FIG. It is a block diagram explaining the structure of the function implement | achieved by the affiliate server 12 which executes a program. It is a figure which shows the example of the collection of queries. It is a figure which shows the example of a standard dictionary. It is a figure which shows the example of a merge dictionary. It is a flowchart explaining the process of a merge dictionary production | generation. It is a flowchart explaining the process of matching text and advertisement. It is a figure which shows the example of the analysis object sentence which is the text used as the object of morphological analysis, and the result of the morphological analysis. It is a figure which shows the example of the result of a morphological analysis. It is a figure explaining acquisition of the advertisement which matches the extracted feature word. It is a figure explaining acquisition of the advertisement which matches the extracted feature word. It is a figure explaining acquisition of the advertisement which matches the extracted feature word. It is a flowchart explaining the other process of matching of a text and an advertisement. It is a flowchart explaining the process of issuance of Javascript. It is a figure which shows the example of Javascript and an advertisement. It is a flowchart explaining the process of an affiliate page display. It is a figure explaining the effect | action of content match delivery API. It is a flowchart explaining the process of purchasing goods. It is a flowchart explaining the process of purchasing goods.

Explanation of symbols

  11 Web server, 12 Affiliate server, 13 Product database, 14 Advertising database, 15 LAN, 16 Router, 17 Network, 18-1, 18-2, or 18 Terminal device, 19-1, 19-2, or 19 User terminal Device, 20-1, 20-2, or 20 user server, 51 CPU, 52 ROM, 53 RAM, 58 storage unit, 61 removable media, 81 merge dictionary generation unit, 82 query set storage unit, 83 standard dictionary storage unit, 84 merge dictionary storage unit, 85 advertisement association unit, 86 association advertisement storage unit, 87 script issuing unit, 88 advertisement providing unit, 89 reward processing unit, 91 generation cost calculation unit, 92 dictionary merge unit, 93 sentence acquisition unit, 94 Analysis Department, 95 Feature Termination Part, 96 ad creation unit

Claims (16)

The product obtained by multiplying the number of times the query for searching the page is used for the search and the number of the search results, which is the number of the searched products, and multiplying each of the plurality of the queries. And calculate a mean value and a standard deviation of the target, and apply a predetermined calculation to the target value and standard deviation target value of the predetermined average value, the average value and standard deviation of the product, and the product for each of the queries Calculating means for calculating a setting value set as a value indicating the likelihood of appearance in each sentence of the query ,
An information processing apparatus comprising: a generation unit that generates a query reflection dictionary by merging the set of queries in which the setting values are calculated with a standard dictionary including a value indicating a universal appearance probability for each word . The information processing apparatus according to claim 1,
The generation means merges the set of queries in which the setting values are calculated into the standard dictionary including a value that is an occurrence cost and that indicates a universal appearance probability for each word. An information processing apparatus comprising: generating the query reflection dictionary. The information processing apparatus according to claim 2,
The information processing apparatus generates the query reflection dictionary by rewriting the occurrence cost of a word in the standard dictionary to the set value of the query that is the same as the word. The information processing apparatus according to claim 1 ,
The calculating means includes
The product C of the number of times used in the search and the number of products searched, the average value AVE of the product C, the standard deviation SD of the product C, the target value NAVE of the average value, and the target value NSD of the standard deviation
An information processing apparatus that calculates the set value for each of the queries by applying an operation represented by an arithmetic expression (product C−average value AVE) / standard deviation SD / 3 × target value NSD + target value NAVE. The information processing apparatus according to claim 1,
An information processing apparatus further comprising extraction means for extracting a feature word from a sentence on a page using the query reflection dictionary. The information processing apparatus according to claim 5 ,
The information processing apparatus extracts the query preferentially as the feature word. The information processing apparatus according to claim 5 ,
An information processing apparatus, further comprising: an acquisition unit configured to acquire an advertisement for a product or service having a feature similar to the feature indicated by the feature word extracted from the text on the page. The information processing apparatus according to claim 7 ,
The acquisition unit acquires an advertisement for a product or a service having a feature similar to the feature indicated by the feature word by searching for the feature word. In an information processing method by an information processing apparatus including a calculation unit and a generation unit,
The calculation means multiplies the number of times a query for searching a page is used for the search by the number of search results, which is the number of products searched, and performs multiplication for each of the plurality of queries. Calculate the average value and standard deviation of the product obtained as a result, and predetermine the target value and standard deviation target value of the predetermined average value, the average value and standard deviation of the product, and the product for each of the queries By applying the calculation, a setting value that is set as a value indicating the likelihood of appearance in each sentence of the query is calculated,
Including a step of generating a query reflection dictionary by merging the set of queries in which the setting values are calculated with a standard dictionary including a value indicating a universal appearance probability for each word by the generation unit. Information processing method. The product obtained by multiplying the number of times the query for searching the page is used for the search and the number of the search results, which is the number of the searched products, and multiplying each of the plurality of the queries. And calculate a mean value and a standard deviation of the target, and apply a predetermined calculation to the target value and standard deviation target value of the predetermined average value, the average value and standard deviation of the product, and the product for each of the queries By calculating a setting value set as a value indicating the likelihood of appearance in each sentence of the query ,
A process including a step of generating a query reflection dictionary by merging the set of queries in which the set values are calculated with a standard dictionary including a value indicating a universal appearance probability for each word is performed on a computer. Program. The product obtained as a result of multiplication for each of the plurality of queries is obtained by multiplying the number of times the query for searching the page is used for the search by the number of the search results, which is the number of products searched. Mean value and standard deviation are calculated, and a predetermined calculation is applied to a predetermined average target value and standard deviation target value, a product average value and standard deviation, and a product for each of the queries. A set of queries for which a set value set as a value indicating the likelihood of appearance in each sentence of the query is calculated, a standard including a value indicating a universal likelihood of appearance for each word An extraction means for extracting a feature word from a sentence on a page using a query reflection dictionary generated by merging with a dictionary;
An information processing apparatus comprising: an acquisition unit configured to acquire an advertisement for a product or a service having characteristics similar to the characteristics indicated by the characteristic words extracted from the text on the page. The information processing apparatus according to claim 11 ,
The information processing apparatus extracts the query preferentially as the feature word. The information processing apparatus according to claim 11 ,
The acquisition unit acquires an advertisement for a product or a service having a feature similar to the feature indicated by the feature word by searching for the feature word. In an information processing method by an information processing apparatus including an extraction unit and an acquisition unit,
The extraction means multiplies the number of times a query for searching a page is used for the search by the number of search results, which is the number of products searched, and performs multiplication for each of the plurality of queries. The average value and standard deviation of the resulting product are calculated, and predetermined values for the target value and standard deviation of the predetermined average value, the average value and standard deviation of the product, and the product for each of the queries are predetermined. Is applied to the set of queries for which a set value set as a value indicating the likelihood of appearance in each sentence of the query is calculated. Extract feature words from the text on the page using a query reflection dictionary generated by merging into a standard dictionary containing the indicated values,
An information processing method including the step of acquiring an advertisement for a product or service having the same characteristics as the characteristics indicated by the characteristic words extracted from the text of the page by the acquisition means . The product obtained as a result of multiplication for each of the plurality of queries is obtained by multiplying the number of times the query for searching the page is used for the search by the number of the search results, which is the number of products searched. Mean value and standard deviation are calculated, and a predetermined calculation is applied to a predetermined average target value and standard deviation target value, a product average value and standard deviation, and a product for each of the queries. A set of queries for which a set value set as a value indicating the likelihood of appearance in each sentence of the query is calculated, a standard including a value indicating a universal likelihood of appearance for each word Using the query reflection dictionary generated by merging with the dictionary, extract feature words from the text on the page,
A program for causing a computer to perform a process including a step of acquiring an advertisement for a product or service having a characteristic similar to the characteristic indicated by the characteristic word extracted from the text on the page. The product obtained by multiplying the number of times the query for searching the page is used for the search and the number of the search results, which is the number of the searched products, and multiplying each of the plurality of the queries. And calculate a mean value and a standard deviation of the target, and apply a predetermined calculation to the target value and standard deviation target value of the predetermined average value, the average value and standard deviation of the product, and the product for each of the queries Calculating means for calculating a setting value set as a value indicating the likelihood of appearance in each sentence of the query ,
Generating means for generating a query reflection dictionary by merging the set of queries in which the setting values are calculated into a standard dictionary including a value indicating a universal appearance probability for each word;
Using the query reflection dictionary, extracting means for extracting feature words from the text on the page;
An information processing system comprising: an acquisition unit configured to acquire an advertisement for a product or a service having characteristics similar to the characteristics indicated by the characteristic words extracted from the text on the page.

Images