JP4311062B2 - Content recommendation system - Google Patents

Description

・・・（式１）
【００９１】
この（式１）においてｎは該当メールマガジンの読者数である。それぞれの読者ｉと，現在の検索者との過去の検索式の一致数がｘ_ｉである。関数ｆ（）は，ｘ_ｉをパラメータとし，パラメータに対し単調増加する性質を持つ関数である。この（式１）の値Ｐは，同じような検索式で今まで検索した人が，数多く該当メールマガジンの読者であるほど数値が高くなる。出力の後，今回の検索式をコンテンツ取得履歴保持手段５に保持（Ｓ７０６）させて，本動作は終了する。
【００９２】
コンテンツ取得履歴リフレッシュ時の動作は，図７を参照しながら説明した上記第２の実施の形態と実質的に同様であるため，重複説明を省略する。
【００９３】
（第３の実施の形態の効果）
以上説明したように，本実施の形態によれば，上記第１の実施の形態または第２の実施の形態と実質的に同様の効果が得られる。さらにそれに加え，本実施の形態ではコンテンツ提供者による登録を必要としないので，例えば本システムを社内のイントラネット上に置く場合など，コンテンツ提供者の協力を得ることが難しい場合でも，実施することができる。
【００９４】
以上，添付図面を参照しながら本発明にかかるコンテンツ推薦システムの好適な実施形態について説明したが，本発明はかかる例に限定されない。当業者であれば，特許請求の範囲に記載された技術的思想の範疇内において各種の変更例または修正例に想到し得ることは明らかであり，それらについても当然に本発明の技術的範囲に属するものと了解される。
【００９５】
例えば，本システムの各手段を構成するハードウェア資源は，各手段でそれぞれ独立している必要はない。例えば，コンテンツ取得履歴保持手段５とコンテンツ情報保持手段６を構成する記憶デバイスは同じ一つのハードディスクであってもよいし，対利用者入力手段２と，コンテンツ情報登録手段３と，コンテンツ要求報告手段４と，コンテンツ取得履歴手段６と，対利用者出力手段７を構成する入力デバイス及び出力デバイスが，全て同じ一つのネットワークカードであってもよい。また逆に，一つの手段に対するハードウェア資源が一つである必要はない。例えば対利用者入力手段は，ある場面ではキーボードで，別の場面ではマウスであってもよい。
【００９６】
本システムの各手段，各手段を構成する各部分，各部分のさらに一部は必ずしも一箇所にある必要はなく，有線または無線ネットワークで接続された範囲で分散配置されてもよい。
【００９７】
【発明の効果】
以上説明したように，本発明によれば，推薦人を用意することなく，利用者の興味や目的に合致する良質な情報を推薦することができる。本発明は，コンテンツの取得とコンテンツ取得解除の動作の両方が生じるコンテンツ，例えばメールマガジンや，その他の会員制で配布されるコンテンツに有効であるほか，コンテンツ取得解除を必要としないコンテンツ，例えば一般に公開されている動画ストリーム等にも有効である。また，社内のイントラネット上に置く場合など，コンテンツ提供者の協力を得ることが難しい場合でも，実施することができる。
【図面の簡単な説明】
【図１】第１の実施の形態にかかるコンテンツ推薦システムの説明図である。
【図２】コンテンツ登録時の動作を示す説明図である。
【図３】コンテンツ検索及び取得時の動作を示す説明図である。
【図４】コンテンツ取得解除時の動作を示す説明図である。
【図５】コンテンツ取得履歴報告時の動作を示す説明図である。
【図６】第１の実施の形態にかかるコンテンツ推薦システムの説明図である。
【図７】コンテンツ取得履歴リフレッシュ時の動作を示す説明図である。
【図８】第３の実施の形態にかかるコンテンツ推薦システムの説明図である。
【図９】コンテンツ登録時の動作を示す説明図である。
【図１０】コンテンツ検索時の動作を示す説明図である。
【符号の説明】
１ 制御手段
２ 対利用者入力手段
３ コンテンツ情報登録手段
４ コンテンツ要求報告手段
５ コンテンツ取得履歴保持手段
６ コンテンツ情報保持手段
７ 対利用者出力手段
８ コンテンツ取得履歴報告手段
１０１ 利用者認証利用者識別子算出部
１０２ コンテンツ取得要求処理部
１０３ コンテンツ取得解除要求処理部
１０４ コンテンツ検索処理部
１０５ コンテンツ情報登録処理部
１０６ コンテンツ取得履歴報告処理部
２０１ コンテンツ取得履歴リフレッシュ処理部
３０１ 受信コンテンツ解析登録処理部[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a content recommendation system that recommends high-quality content that matches a user's interest and purpose from various data (content) such as a large amount of documents, sounds, and images distributed via a network.
[0002]
[Prior art]
The spread of the Internet in recent years has made it possible to distribute and view various data (contents) such as a large amount of documents, sounds, and images. This expansion has made it easier for people to obtain a variety of content, but makes it difficult to select the desired high-quality content. As a means for solving this problem, a method of recommending content has been considered (for example, JP-A-2002-42333, JP-A-2002-342370, etc.). These are intended to allow the user to acquire the desired high-quality content by setting the recommender and presenting the content recommended by the recommender to the user.
[0003]
[Patent Document 1]
JP 2002-43233 A
[Patent Document 2]
JP 2002-342370 A
[0004]
[Problems to be solved by the invention]
However, methods such as Japanese Patent Laid-Open No. 2001-43233 and Japanese Patent Laid-Open No. 2002-342370 need to prepare a recommender who is a person, so that content without a recommender is naturally not recommended. For those who have no interest in the content provider, it is troublesome to perform recommendation work, but there is no reward for it, so it is difficult to secure many such recommenders. In other words, a person who becomes a content recommender tends to recommend a content that has a stake in a content provider. As a result, such methods are likely to have little or no recommendation.
[0005]
The present invention has been made in view of the above problems of the conventional content recommendation system, and the object of the present invention is to provide high-quality information that matches the interests and purposes of the user without preparing a recommender. It is to provide a new and improved content recommendation system that can be recommended.
[0006]
[Means for Solving the Problems]
  In order to solve the above problems, according to a first aspect of the present invention, there is provided a system for recommending content, and content acquisition that retains a content acquisition history in which a search condition used for content acquisition is associated with the content A history holding means (5), a content search processing unit (104) that refers to the content acquisition history and recommends content related to the search condition from the user in response to a content acquisition request based on the search condition from the user; A content recommendation system is provided. Then, when there are a plurality of recommended contents, the content search processing unit (104) refers to the content acquisition history, and according to the number of contents acquired using the search conditions related to the search conditions from the user. And determining the recommendation order of content.The

[0007]
According to such a system, when there are a plurality of recommended content, the content search processing unit (104) refers to the content acquisition history and acquires the content using the search condition related to the search condition from the user. Since the recommendation order of contents is determined according to the number of cases, it is possible to recommend high-quality information that matches the user's interests and purpose without preparing a recommender.
[0008]
The content recommendation system according to the first aspect of the present invention can be applied as follows.
[0009]
  The content search processing unit (104) can increase the recommendation order of content as the number of contents acquired using the search condition related to the search condition from the user increases.The

[0010]
  The search condition related to the search condition from the user can be a search condition that matches the search condition from the user.The
[0011]
  Content acquisition history can include information about the user who acquired the content.TheIn this case, the information about the user who acquired the content can include a code obtained by irreversibly converting the code representing the user.TheEven if the data contents are intruded into the system for some reason, the search conditions of each user cannot be specified and the privacy problem can be solved.
[0012]
  It is possible to include a content acquisition request processing unit (102) that performs a content acquisition history update process in response to a content acquisition request from a user.The
[0013]
  It is possible to include a content acquisition cancellation request processing unit (103) that performs a content acquisition history update process in response to a content acquisition cancellation request from a user.TheThis is particularly effective for content that undergoes both content acquisition and content acquisition cancellation operations, such as e-mail magazines and other content distributed under membership.
[0014]
  A content acquisition history refresh processing unit (201) that performs a content acquisition history update process when a predetermined update condition is satisfied regardless of a request from a user (content acquisition request or content acquisition release request, etc.) It is possible to includeTheThis is particularly effective for content that does not require content acquisition cancellation, such as a publicly available video stream.
[0015]
  In this case, the content acquisition history refresh processing unit (201) can delete the content acquisition history older than the time specified by the user.TheAlternatively, the content acquisition history refresh processing unit (201) can also delete a content acquisition history having a smaller number of referenced times than the number of referenced times specified by the user.The
[0016]
  In order to solve the above problem, according to a second aspect of the present invention,A system for recommending content, in which a search condition used for searching for content is associated with a content search history associated with a searcher who has searched the content, and content is associated with a provider who has transmitted the content to the system. Content acquisition history holding means (5) for holding the acquired content acquisition history, and in response to a content acquisition request based on the search condition from the searcher, the search condition related to the search condition from the searcher is included in the content search history. When the searcher associated with the search condition and the searcher included in the content search history is included as the provider in the content acquisition history, the content associated with the provider is extracted from the content acquisition history. A content search processing unit (104) to be recommended, and a content search processing unit (104) , If the content to be recommended there are multiple, depending on the relevance of the search conditions included in the search condition and content search history from searcher, and determines a recommendation order of the content(Claims1).
[0018]
In the content recommendation system according to the second aspect of the present invention, the following applications are possible.
[0019]
  The content search processing unit (104) is associated with the provider associated with the search condition included in the content search history as the degree of matching between the search condition included in the content search history and the search condition from the searcher increases. It is possible to raise the recommendation order of content(Claims2). And in this case,The degree of coincidence between the search condition included in the content search history and the search condition from the searcher is based on the search condition used for the search included in the content search history and the search condition from the searcher.Can be given by a monotonically increasing function as a parameter.3).
[0020]
  SearchSearch conditions related to search conditions from theSearchSearch conditions that match the search conditions from the user (claims)4).
[0021]
  Content acquisition history acquired contentOfferCan include information about the person (claims)5). And in this case, the content was acquiredOfferFor information on the personOfferIt is possible to include a code obtained by irreversibly transforming a code representing a person (claims).6). Even if the data contents are intruded into the system for some reason, the search conditions of each user cannot be specified and the privacy problem can be solved.
[0022]
  further,OfferThe content receiving means (9) for receiving the content acquired by the user and the content received by the content receiving means (9), at least the content of the content and the content transmittedOfferAnd a received content analysis registration processing unit (301) for registering in the content acquisition history holding means.7).
[0023]
  According to another aspect of the present invention, a computer program for causing a computer to operate (function) as the content recommendation system according to the first or second aspect described above (claims)8And a computer-readable recording medium on which the computer program is recorded (claims)9) Is provided. Here, the program may be described in any programming language. As the recording medium, for example, CD-ROM, DVD-ROM, floppy disk(Registered trademark)Disk (FD: Floppy(Registered trademark)  As a recording medium capable of recording a program, such as a disk), a recording medium generally used at present or any recording medium used in the future can be adopted.
[0024]
In addition, in the above, in order to facilitate understanding, the constituent elements described in parentheses are merely the corresponding constituent elements and signals in the embodiments described later, and the present invention is not limited thereto.
[0025]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, preferred embodiments of a content recommendation system according to the present invention will be described in detail with reference to the accompanying drawings. In the present specification and drawings, components having substantially the same functional configuration are denoted by the same reference numerals, and redundant description is omitted.
[0026]
(First embodiment)
FIG. 1 is an explanatory diagram of a content recommendation system (hereinafter also simply referred to as the present system) according to the present embodiment. As shown in FIG. 1, the content recommendation system 10 according to the present embodiment includes a control unit 1, a user input unit 2, a content information registration unit 3, a content request report unit 4, and a content acquisition history. The holding means 5, the content information holding means 6, the user output means 7, and the content acquisition history report means 8 are configured.
[0027]
The control means 1 controls the following 2 to 8 means according to a predetermined procedure, and may be configured on hardware resources including a CPU (Central Processing Unit) and a memory, A dedicated circuit may be used. The detailed configuration of the control means 1 will be further described later.
[0028]
The user input means 2 is a means for the user of this system to input conditions for searching for contents and the like, and is composed of hardware resources including an input device such as a keyboard, mouse, or network card.
[0029]
The content information registration means 3 is a means for the content provider to register a keyword, an outline, or the content itself representing the content to be provided, and the procedures for acquiring and canceling the content in this system. Or hardware resources including an input device such as a network card.
[0030]
The content request reporting unit 4 is a unit for reporting to the content provider that the user has requested acquisition or cancellation of the content, and output such as a CRT (Cathode Ray Tube), a printer, or a network card. Consists of hardware resources including devices.
[0031]
The content acquisition history holding means 5 is means for holding a request history for acquiring contents performed by the user through this system, and includes a hard disk, a magneto-optical disk, a hard disk including a storage device such as NVRAM (Non Volatile RAM). Hardware resources.
[0032]
The content information holding unit 6 is a unit for holding the keyword representing the content, the outline, or the content itself, and the procedures for acquiring and canceling the content input by the content information registering unit 3. It is composed of hardware resources including storage devices such as disks and NVRAM (Non Volatile RAM).
[0033]
The user output means 7 is a means for outputting a response to the input made by the user input means 2, and is composed of hardware resources including an output device such as a CRT, a printer or a network card.
[0034]
The content acquisition history reporting means 8 is means for reporting the content held in the content acquisition history holding means 5 to the content provider, and is constituted by hardware resources including an output device such as a CRT, a printer, or a network card. .
[0035]
The configuration of the control means 1 will be described in detail.
The control means 1 includes a user authentication user identifier calculation unit 101, a content acquisition request processing unit 102, a content acquisition cancellation request processing unit 103, a content search processing unit 104, a content information registration processing unit 105, and a content acquisition. A history report processing unit 106 is provided.
[0036]
The user authentication user identifier calculation unit 101 authenticates a user using data input from the user input means 2 and calculates a user identifier indicating the user. This user identifier is preferably irreversible.
[0037]
The content acquisition request processing unit 102 uses the content request reporting unit 4 in accordance with the content acquisition procedure held in the content information holding unit 6 in response to the content acquisition request input from the user input unit 2. The corresponding content provider is notified that the user is requesting acquisition of the corresponding content, and the history (user identifier, search condition, and acquired content identifier) is stored in the content acquisition history holding means 5.
[0038]
The content acquisition release request processing unit 103 responds to the content acquisition release request input from the user input means 2 according to the content acquisition release procedure held in the content information holding means 6 and the content request reporting means. 4 is used to report to the corresponding content provider that the user has requested to cancel the acquisition of the corresponding content, and the previously stored history (user identifier, search condition, and acquired content identifier) is acquired as content. Delete from the history holding means 5.
[0039]
The content search processing unit 104 searches the content information holding unit 6 for content that satisfies the search condition input by the user input unit 2, searches the content acquisition history holding unit 5 for each searched content, The number of cases searched under the same search condition is obtained, the recommendation order is determined in descending order of the number of cases, and the result is output by the user output means 7.
[0040]
The content information registration processing unit 105 stores, in the content information holding unit 6, the keywords representing each content, the outline, or the content itself input by the content information registration unit 3, and the content acquisition and acquisition request procedures. , Assign a unique content identifier to each content.
[0041]
Content identifier assignment will be described.
A content identifier is assigned to each content. A mail magazine will be described as an example of content. For example, if there is a mail magazine “Oki Mail Magazine March 17th issue”, one content identifier is assigned to it. By linking the content identifier and the user identifier, the user of “Oki Mail Magazine March 17 issue” can be searched.
[0042]
This content identifier needs to be unique not only for one content provider but also for other content providers. However, since the content identifier is assigned on the system side, the content provider need not be aware of the assignment of the content identifier.
[0043]
The simplest means for assigning content identifiers is simply a serial number. The content to be registered is compared with the content registered so far, and if there is no match, a new number may be assigned. This is sufficient for non-continuous content.
[0044]
On the other hand, for example, the following content identifier assignment method can be considered for continuous content such as a mail magazine issued every day. That is, first, the type of the mail magazine is determined by matching the mail header and the fixed part unique to the mail magazine. Then, a part that recognizes the type of the mail magazine and a part that recognizes each issue of the mail magazine can be constituted by a content identifier. According to such a content identifier assignment method, the process of collecting mail magazines is simplified. This is because in the case of an e-mail magazine, users who receive e-mail magazines receive not only the issue but also the subsequent issues, so it is desirable to process the entire e-mail magazine rather than each issue individually. .
[0045]
The configuration of the control unit 1 will be described again. The content acquisition history report processing unit 106 reports the statistical information of the history stored in the content acquisition history holding unit 5 with the content acquisition history report unit 8.
[0046]
The content recommendation system 10 according to the present embodiment is configured as described above. Next, the operation of the content recommendation system 10 will be described. Note that there are various contents such as documents, images, and audio that can be applied in the present embodiment. Here, in order to facilitate understanding, a case where a mail magazine is applied will be described as one specific example.
[0047]
(Operation of the first embodiment)
FIG. 2 shows the flow of operations when registering a mail magazine (content). An e-mail magazine provider who wishes to recommend e-mail magazines using this system uses the content information registration means 3 to provide an overview of e-mail magazines to be searched for by the user, keywords indicating characteristics, or e-mail magazines. And the mail magazine acquisition and acquisition cancellation procedure are input (S101).
[0048]
The acquisition and acquisition cancellation procedure here is a group of data that can be understood by the control unit 1 when the system uses the content request reporting unit 4 to notify the mail magazine provider of an acquisition and acquisition cancellation request. Specifically, an e-mail address or URL (Unified Resource Locator) indicating the e-mail magazine application destination, e-mail address to which the e-mail magazine should be sent, personal information to be registered, and notification with the e-mail magazine issuer Procedures. The input data group is stored in the content information holding means 6 (S102).
[0049]
FIG. 3 shows a flow of operations at the time of mail magazine (content) search and acquisition. When a user of this system tries to search for an e-mail magazine, the user first performs authentication by entering the user name notified to the user in advance using the user input means 2 and a password if necessary. The user identifier used in the following processing is calculated (S201).
[0050]
The algorithm for generating user identifiers is arbitrary, but at least individual user identifiers must be generated for each user. Furthermore, this user identifier is preferably irreversible. In other words, an individual user identifier can be uniquely generated from the user name, but the user cannot be immediately identified from the user identifier. In this case, the user identifier should be calculated for each authentication, and a correspondence table between the converted user name and the user identifier should not be created and stored in a storage medium capable of long-term storage.
[0051]
When the user authentication and the calculation of the user identifier are completed, the user inputs a search condition using the user input means 2 (S202). This may be a form of selecting from a list prepared in advance, or a form of inputting an arbitrary word or natural sentence. However, when adopting the form of inputting arbitrary single word sounds or natural sentences, it is desirable to add a process that corrects differences in the wording and expression method of each input.
[0052]
As an example of a process for correcting the difference between individuals, there is a simple one-byte character / single-byte character unification. For example, the word “IP” is written in full-width characters or written in half-width characters by the user, so it is unified with either. As another example, it is conceivable to unify the wording using, for example, a synonym dictionary.
[0053]
When the search condition is input, the control unit 1 searches the content information holding unit 6 for a mail magazine satisfying the search condition (S203), and further searches the content acquisition history holding unit 5 for each searched mail magazine. Then, the number of cases searched under the same condition is obtained (S204), the recommendation order is determined in the order of the largest number, and the result is output by the user output means 7 (S205).
[0054]
This recommends to users a mail magazine with a large number of readers who have already searched with the same search conditions from among mail magazines that satisfy the search conditions. The fact that there are a large number of readers who search for an e-mail magazine using the same search conditions means that the e-mail magazine matches the interests and purpose of the user who searches for the e-mail magazine using the search conditions and is of good quality. It is a thing.
[0055]
A system that simply recommends a search that matches the search criteria does not judge the quality of the content. Also, if you simply count on the number of readers in an email magazine, the email magazine may have been popular in other fields. However, according to this system, such a problem can be avoided.
[0056]
From the recommendation result output from the user output means 7, the user makes an acquisition request using the user input means 2 if there is a desired mail magazine (S206). When there is an acquisition request, according to the acquisition procedure stored in step S102, the request is notified to the mail magazine provider using the content request reporting means 4 (S207), and further, the acquisition history of the mail magazine (user identifier) And the search condition and the identifier of the acquired mail magazine) are stored in the content acquisition history holding means 5 (S208). This memory is used in step S204 by other users.
[0057]
FIG. 4 shows the flow of operation when the e-mail magazine (content) acquisition is canceled. When the user of this system tries to cancel the acquisition of the e-mail magazine, first, the user input means 2 is used to perform the authentication work by inputting the user name and the password if necessary, and the process of S201. A user identifier is calculated using the same algorithm as in step (S301). Next, the e-mail magazine acquisition history that matches the user identifier is searched from the content acquisition history holding means 5 (S302). This is the own mail magazine acquisition history, and this result is output using the user output means 7 (S303).
[0058]
The user selects an unnecessary mail magazine from the output, and inputs it using the user input means 2 (S304). When there is a request for cancellation of acquisition, the request is notified to the mail magazine provider using the content request reporting means 4 according to the acquisition cancellation procedure stored in the step of S102 (S305), and the acquisition history of the mail magazine ( The user identifier, the search condition, and the acquired mail magazine identifier) are deleted from the content acquisition history holding means 5 (S306).
[0059]
FIG. 5 shows a flow of operations for reporting a mail magazine (content) acquisition history to a content provider. This searches the content acquisition history holding means 5 for each mail magazine to acquire the number of people and search conditions (S401), and reports it to the mail magazine provider using the content acquisition history report means 8. As a result, the mail magazine provider can know the keyword expected by the reader of each mail magazine.
[0060]
(Effects of the first embodiment)
As described above, according to the present embodiment, it is possible to recommend high-quality information that matches the user's interest and purpose without preparing a recommender.
[0061]
In addition, by using the value calculated by irreversible transformation as a user identifier, it is possible to specify the search condition of each user even if the data contents are illegally viewed due to intrusion into the system for some reason. It becomes impossible and can solve privacy problems.
[0062]
This embodiment is particularly effective for content that undergoes both content acquisition and content acquisition cancellation operations, such as e-mail magazines and other content distributed by membership.
[0063]
(Second Embodiment)
FIG. 6 is an explanatory diagram of a content recommendation system (hereinafter also simply referred to as the present system) according to the present embodiment. Only the differences from the content recommendation system 10 according to the first embodiment shown in FIG. 1 will be described. The content recommendation system 20 according to the present embodiment cancels content acquisition as shown in FIG. The request processing unit 103 does not exist, and instead the content acquisition history refresh processing unit 201 exists.
[0064]
The content acquisition history refresh processing unit 201 is activated under a predesignated condition, and deletes a history that has exceeded a predesignated period from the time of registration from the content acquisition history holding means 5 (refresh).
[0065]
The other components are substantially the same as those in the first embodiment, and a duplicate description is omitted.
[0066]
The content recommendation system 20 according to the present embodiment is configured as described above. Next, the operation of the content recommendation system 20 will be described. As in the first embodiment, there are various contents such as documents, images, and audio that can be applied to the present embodiment. Here, in order to facilitate understanding, as one specific example, A case where a video stream is applied will be described.
[0067]
(Operation of Second Embodiment)
FIG. 2 shows the flow of operations when registering a moving image stream (content). A video stream provider who wishes to recommend a video stream using this system uses the content information registration means 3 to provide an overview of the video stream to be searched for by the user, keywords representing characteristics, or a video stream. A snapshot or the like and a moving image stream acquisition procedure are input (S101).
[0068]
The acquisition procedure here is a group of data that can be understood by the control means 1 and is necessary for the system to notify the video stream provider of the acquisition request using the content request reporting means. E-mail address and URL (Unified Resource Locator) indicating the application destination of the stream, e-mail address that should notify the applicant of the viewing method of the video stream (URL, password, etc.), items such as personal information to be registered, and video For example, a notification procedure with the stream provider. The input data group is stored in the content information holding means 6 (S102).
[0069]
FIG. 3 shows a flow of operations when searching and acquiring a moving image stream (content). When a user of this system tries to search for a video stream, first, the user input means 2 is used to perform authentication work by entering a user name and a password as necessary. A user identifier to be used is calculated (S201).
[0070]
The algorithm for generating user identifiers is arbitrary, but at least individual user identifiers must be generated for each user. Furthermore, this user identifier is preferably irreversible. In other words, an individual user identifier can be uniquely generated from the user name of the user, but the user cannot be immediately identified from the user identifier. In this case, the user identifier is calculated every time authentication is performed. A correspondence table between user names and user identifiers should not be created and stored in a storage medium capable of long-term storage.
[0071]
When the user authentication and the calculation of the user identifier are completed, the user next inputs a search condition using the user input means 2 (S202). This may be a form of selecting from a list prepared in advance, or a form of inputting an arbitrary word or natural sentence. However, when adopting the form of inputting arbitrary words and natural sentences, it is desirable to add processing that corrects the differences in the wording and expression method of each input.
[0072]
When the search condition is input, the control unit 1 searches the content information holding unit 6 for a video stream satisfying the search condition (S203), and further, the content acquisition history holding unit 5 is set for each of the searched video streams. A search is made to find the number of cases searched under the same search conditions (S204). The recommendation order is determined in descending order of the number of cases, and the results are output using the user output means 7 (S205).
[0073]
This recommends to the user a video stream with a large number of people who have already been searched and viewed under the same search condition from among the video streams satisfying the search condition. The fact that there are many viewers who have searched and viewed a video stream under the same search conditions means that the video stream matches the interest and purpose and is of good quality for the users who search using the search conditions. It is that.
[0074]
A system that simply recommends a search that matches the search criteria does not judge the quality of the content. In addition, if only the viewer of the video stream is used as a condition, the video stream may have been popular in other topics. However, according to this system, such a problem can be avoided.
[0075]
From the result output from the user output means 7, the user makes an acquisition request using the user input means 2 if there is a desired video stream (S206). When there is an acquisition request, according to the acquisition procedure stored in step S102, the request is notified to the video stream provider using the content request report means 4 (S207), and the acquisition history of the video stream (user identifier) And the search condition and the identifier of the acquired moving image stream) are stored in the content acquisition history holding means 5 (S208). This memory is used in the step S204 of other users.
[0076]
FIG. 5 shows a flow of operations for reporting a moving image stream (content) acquisition history to a content provider. This searches the content acquisition history holding means 5 for each video stream to acquire the number of people and search conditions (S401), and reports it to the video stream provider using the content acquisition history report means 8. Thereby, the video stream provider can know the keyword expected by each video stream viewer.
[0077]
FIG. 7 shows the flow of operations when refreshing the content acquisition history. The content acquisition history refresh is started in a predetermined condition, for example, at a specified time interval or when the number of histories stored in the content acquisition history holding means 5 exceeds a specified number. The content acquisition history is deleted in a predetermined order in the history stored in the content acquisition history holding means 5, for example, in order from the oldest, or in order from the smallest reference count (referenced count). , Refresh (S501).
[0078]
(Effect of the second embodiment)
As described above, according to the present embodiment, it is possible to recommend information of good quality that matches the user's interest and purpose without preparing a recommender.
[0079]
In addition, by using the value calculated by irreversible transformation as a user identifier, it is possible to specify the search condition of each user even if the data contents are illegally viewed due to intrusion into the system for some reason. It becomes impossible and can solve privacy problems.
[0080]
This embodiment is particularly effective for content that does not require content acquisition cancellation, for example, a publicly available video stream.
[0081]
(Third embodiment)
FIG. 8 is an explanatory diagram of a content recommendation system (hereinafter also simply referred to as this system) according to the present embodiment. Only the differences from the content recommendation system 20 according to the second embodiment shown in FIG. 6 will be described. The content recommendation system 30 according to the present embodiment performs content information registration as shown in FIG. The means 3, the content requesting means 4, the content acquisition request processing unit 102, and the content information registration unit 105 do not exist. Instead, there are a content receiving unit 9, a content acquisition history request unit 10, and a received content analysis registration processing unit 301.
[0082]
The content receiving means 9 is means for receiving the content acquired by the user, and is composed of hardware resources including an input device such as a network card. The content acquisition history request unit 10 is a unit for requesting that the content provider output the content held in the content acquisition history holding unit 5 to the content acquisition history report unit 8, and is a keyboard, mouse, or Consists of hardware resources including input devices such as network cards.
[0083]
The received content analysis / registration processing unit 301 analyzes the content of the content received by the content receiving unit 9 and holds it in the content information holding unit 6, determines the sender of the content, and holds it in the content acquisition history holding unit Let
[0084]
The other components are substantially the same as those in the first or second embodiment, and thus the duplicate description is omitted.
[0085]
The content recommendation system 30 according to the present embodiment is configured as described above. Next, the operation of the content recommendation system 30 will be described. As in the first or second embodiment, there are various contents such as a document, an image, and a sound that can be applied to the present embodiment. Here, in order to facilitate understanding, one specific example is used. As an example, a case in which a mail magazine is applied will be described as in the first embodiment.
[0086]
(Operation of the third embodiment)
FIG. 9 shows the flow of operations when registering a mail magazine (content). Unlike the first and second embodiments, the mail magazine provider does not need to register information in this embodiment. Instead, each user sends the e-mail magazine itself already acquired to the system. The system receives these mail magazines using the content receiving means 9 (S601).
[0087]
Since multiple e-mail magazines receive the same content, the same e-mail magazine content is sent to the system multiple times. Therefore, it is checked whether or not the sent mail magazine has already been sent (S602). If it is sent for the first time, it is stored in the content information holding means 6 (S603). Then, the user of this system is determined from the sender address, the user identifier is calculated by the same algorithm as in step S201 (S604), and the mail magazine acquisition history (user identifier and acquired mail is stored in the content acquisition history holding means 5). Magazine identifier) is stored (S605).
[0088]
FIG. 10 shows the flow of operations when retrieving a mail magazine (content). Steps S701, S702, and S703 are substantially the same as steps S201, S202, and S203 shown in FIG. Next to step S703, for each of the mail magazines searched in step S703, the contents of the history that other users acquired the mail magazine in the past and the search formulas that have been input so far by other users are acquired. Obtained from the history holding means 7 (S704).
[0089]
Then, the corresponding mail magazine is ranked so that the one with a high P in the following (formula 1) is ranked higher, and the result is output (S705).
[0090]
[Expression 1]

... (Formula 1)
[0091]
In this (Formula 1), n is the number of readers of the corresponding mail magazine. The number of matches in the past search formula between each reader i and the current searcher is x_iIt is. The function f () is x_iIs a function that has a property of monotonically increasing with respect to the parameter. The value P of (Equation 1) increases as the number of people who have searched with the same search formula so far increases the number of readers of the corresponding e-mail magazine. After the output, the current retrieval formula is held in the content acquisition history holding means 5 (S706), and this operation is finished.
[0092]
Since the operation at the time of content acquisition history refresh is substantially the same as that of the second embodiment described with reference to FIG.
[0093]
(Effect of the third embodiment)
As described above, according to the present embodiment, substantially the same effect as the first embodiment or the second embodiment can be obtained. In addition to this, since this embodiment does not require registration by the content provider, it can be performed even when it is difficult to obtain cooperation from the content provider, for example, when the system is placed on an intranet in the company. it can.
[0094]
The preferred embodiment of the content recommendation system according to the present invention has been described above with reference to the accompanying drawings, but the present invention is not limited to such an example. It will be obvious to those skilled in the art that various changes or modifications can be conceived within the scope of the technical idea described in the claims, and these are naturally within the technical scope of the present invention. It is understood that it belongs.
[0095]
For example, the hardware resources constituting each means of this system need not be independent from each other. For example, the storage devices constituting the content acquisition history holding means 5 and the content information holding means 6 may be the same hard disk, the user input means 2, the content information registration means 3, and the content request reporting means. 4, the content acquisition history unit 6, and the input device and output device that constitute the user output unit 7 may all be the same network card. Conversely, there is no need for one hardware resource for one means. For example, the user input means may be a keyboard in one scene and a mouse in another scene.
[0096]
Each means of the present system, each part constituting each means, and further part of each part do not necessarily have to be in one place, and may be distributed in a range connected by a wired or wireless network.
[0097]
【The invention's effect】
As described above, according to the present invention, it is possible to recommend high-quality information that matches the user's interest and purpose without preparing a recommender. The present invention is effective for content in which both content acquisition and content acquisition cancellation operations occur, such as e-mail magazines and other content distributed under a membership system, and content that does not require content acquisition cancellation, for example, in general It is also effective for publicly available video streams. It can also be implemented even when it is difficult to obtain cooperation from the content provider, such as when it is placed on the company intranet.
[Brief description of the drawings]
FIG. 1 is an explanatory diagram of a content recommendation system according to a first embodiment.
FIG. 2 is an explanatory diagram showing an operation at the time of content registration.
FIG. 3 is an explanatory diagram showing operations during content search and acquisition.
FIG. 4 is an explanatory diagram showing an operation when canceling content acquisition;
FIG. 5 is an explanatory diagram showing an operation when reporting a content acquisition history.
FIG. 6 is an explanatory diagram of a content recommendation system according to the first embodiment.
FIG. 7 is an explanatory diagram showing an operation at the time of content acquisition history refresh.
FIG. 8 is an explanatory diagram of a content recommendation system according to a third embodiment.
FIG. 9 is an explanatory diagram showing an operation at the time of content registration.
FIG. 10 is an explanatory diagram showing an operation during content search.
[Explanation of symbols]
1 Control means
2 User input means
3 Content information registration means
4 Content request reporting means
5 Content acquisition history holding means
6 Content information holding means
7 User output means
8 Content acquisition history reporting means
101 User authentication user identifier calculation unit
102 Content acquisition request processing unit
103 Content acquisition release request processing unit
104 Content search processing unit
105 Content information registration processing unit
106 Content acquisition history report processing unit
201 Content acquisition history refresh processing unit
301 Received content analysis registration processing unit

Claims (9)

  A system for recommending content,
  Content that retains a content search history that associates the search condition used for content retrieval with the searcher who searched the content, and a content acquisition history that associates the content with the provider who transmitted the content to the system. Acquisition history holding means;
  In response to a content acquisition request from a searcher, a search condition associated with the search condition from the searcher is included in the content search history, the searcher is associated with the search condition, and the searcher A content search processing unit that recommends the content associated with the provider from the content acquisition history when a searcher included in the content search history is included as the provider in the content acquisition history;
  The content search processing unit, when there are a plurality of recommended contents, determines a content recommendation order according to a relationship between a search condition from the searcher and a search condition included in the content search history. Content recommendation system characterized by The content search processing unit matches the searcher associated with the search condition included in the content search history as the degree of match between the search condition included in the content search history and the search condition from the searcher increases. The content recommendation system according to claim 1 , wherein a recommendation ranking of the content associated with the provider is increased. The degree of matching between the search condition included in the content search history and the search condition from the searcher is:
A search condition used in the search included in the content search history, characterized in that provided by monotonically increasing function degree of coincidence between retrieval conditions from the searcher as a parameter, according to claim 2 Content recommendation system. Search condition associated with the search condition from the searcher, characterized in that it is a search condition that matches the search conditions from the searcher, the content recommendation system according to claim 1. The content recommendation system according to any one of claims 1 to 4 , wherein the content acquisition history includes information on a provider who acquired the content. The content recommendation system according to claim 5 , wherein the information related to the provider who acquired the content includes a code obtained by irreversibly converting a code representing the provider . further,
Content receiving means for receiving content acquired by the provider ;
A received content analysis / registration processing unit that analyzes at least the content of the content and the provider that transmitted the content and registers the content in the content acquisition history holding unit with respect to the content received by the content receiving unit; The content recommendation system according to any one of claims 1 to 6 . A computer program for causing a computer to operate as the content recommendation system according to any one of claims 1 to 7 . A computer-readable recording medium on which the computer program according to claim 8 is recorded.

Images