CN104699687A

CN104699687A - Item recommendation method and server

Info

Publication number: CN104699687A
Application number: CN201310648416.XA
Authority: CN
Inventors: 张文
Original assignee: China Mobile Group Guangdong Co Ltd
Current assignee: China Mobile Group Guangdong Co Ltd
Priority date: 2013-12-04
Filing date: 2013-12-04
Publication date: 2015-06-10

Abstract

The invention provides an item recommendation method and a server. The item recommendation method comprises the following steps: acquiring history use information of items to be recommended, wherein the history use information of the items to be recommended comprises history use position information of the items to be recommended and/or history use time information of the items to be recommended; acquiring current information of a terminal of a user to be recommended, wherein the current information of the terminal of the user to be recommended comprises the current position information and/or current time of the terminal of the user to be recommended; determining an item recommendation list according to the history use information of the items to be recommended and the current information of the terminal of the user to be recommended; sending the item recommendation list to the terminal of the user to be recommended. By virtue of the technical scheme, the items can be recommended according to the history use information of the items without position attribution information.

Description

A kind of item recommendation method and server

Technical field

The present invention relates to Internet of Things field, particularly relate to a kind of item recommendation method and server.

Background technology

Item recommendation method utilizes Information Filtering Technology to recommend the interested information of its possibility to user.Item recommendation method is different from the information processing manner of information classification and information search, is inferred draw by user behavior.In the prior art, item recommendation method is recommended based on the position of article, and namely article must have geographic position attribute.And in the article recommendation process of reality, the article with geographic position attribute and time attribute only account for a part, all article can not be covered.In addition, in the prior art, item recommendation method cannot be recommended based on the service time of article.

Summary of the invention

In order to solve the problem, the invention provides a kind of item recommendation method and server, described item recommendation method does not limit article must have geography information attribute, avoid do not have the article of geography information attribute cannot participate in recommend and cannot based on the problem of carrying out the service time of article recommending.

To achieve these goals, the invention provides a kind of item recommendation method, comprise: the history obtaining article to be recommended uses information, the history of described article to be recommended uses information to comprise the history use location information of article to be recommended and/or history information service time of article to be recommended, obtain the current information of the terminal of user to be recommended, the current information of the terminal of described user to be recommended comprises current location information and/or the current time of the terminal of user to be recommended, and when the history of described article to be recommended uses information to comprise the history use location information of article to be recommended, the current information of the terminal of described user to be recommended comprises the current location information of the terminal of user to be recommended, when the history of described article to be recommended uses information to comprise history information service time of article to be recommended, the current information of the terminal of described user to be recommended comprises current time, use the current information of the terminal of information and described user to be recommended according to the history of described article to be recommended, determine article recommendation list, described article recommendation list is sent to the terminal of described user to be recommended.

Further, the described history according to described article to be recommended uses the current information of the terminal of information and described user to be recommended, before determining article recommendation list, also comprises: the history obtaining user to be recommended uses Item Information; Determine that the history of user to be recommended uses the degree of correlation of article and each article to be recommended; Use the degree of correlation of article and each article to be recommended according to the history of user to be recommended, determine initial article recommendation list; The described history according to described each article to be recommended uses the current information of the terminal of information and described user to be recommended, determine article recommendation list, be specially: the current information using the terminal of information and described user to be recommended according to the history of described initial article recommendation list, described each article to be recommended, determine article recommendation list.

Further, the history of described article to be recommended uses information to comprise the history use location information of article to be recommended, the current information of the terminal of described user to be recommended comprises the current location information of the terminal of user to be recommended, the described history according to described article to be recommended uses the current information of the terminal of information and described user to be recommended, determine article recommendation list, be specially: according to the current location information of the history use location information of described article to be recommended and the terminal of user to be recommended, determine the minimum distance of the current location of the history use location of each article to be recommended and the terminal of user to be recommended, according to the minimum distance of the current location of the described history use location of each article to be recommended and the terminal of user to be recommended, determine article recommendation list.

Further, the history of described article to be recommended uses information to comprise history information service time of article to be recommended, the current information of the terminal of described user to be recommended comprises current time, the described history according to described article to be recommended uses the current information of the terminal of information and described user to be recommended, determine article recommendation list, be specially: according to history information service time and the current time of described article to be recommended, determine the history service time of each article to be recommended and the nearest time difference of current time; According to the described history service time of each article to be recommended and the nearest time difference of current time, determine article recommendation list.

Further, the history of described article to be recommended uses information to comprise the history use location information of article to be recommended and history information service time of article to be recommended, the current information of the terminal of described user to be recommended comprises current location information and the current time of the terminal of user to be recommended, the described history according to described article to be recommended uses the current information of the terminal of information and described user to be recommended, determine article recommendation list, be specially: according to the current location information of the history use location information of described article to be recommended and the terminal of user to be recommended, determine the minimum distance of the current location of the history use location of each article to be recommended and the terminal of user to be recommended, according to history information service time and the current time of described article to be recommended, determine the history service time of each article to be recommended and the nearest time difference of current time, according to the minimum distance of the current location of the described history use location of each article to be recommended and the terminal of user to be recommended and the history service time of each article to be recommended and the nearest time difference of current time, determine article recommendation list.

Further, the described current location information according to the history use location information of described article to be recommended and the terminal of user to be recommended, before determining the minimum distance of the current location of the history use location of each article to be recommended and the terminal of user to be recommended, also comprise: according to history use location information and the clustering algorithm of described article to be recommended, determine the history use location cluster set of each article to be recommended; The described current location information according to the history use location information of described article to be recommended and the terminal of user to be recommended, determine the minimum distance of the current location of the history use location of each article to be recommended and the terminal of user to be recommended, be specially: according to the history use location cluster set of each article to be recommended, determine the cluster centre position of the history use location of each article to be recommended; The minimum distance of the current location of the history use location of each article to be recommended and the terminal of user to be recommended is determined according to the current location of the cluster centre position of the history use location of each article to be recommended and the terminal of user to be recommended.

Further, the minimum distance of the described current location according to the history use location of each article to be recommended and the terminal of user to be recommended, determine article recommendation list, be specially: according to minimum distance and the default distance coefficient of the current location of the described history use location of each article to be recommended and the terminal of user to be recommended, determine the distance score of each article to be recommended; According to the distance score of each article to be recommended, determine article recommendation list.

Further, described history information service time according to described article to be recommended and current time, before determining the history service time of each article to be recommended and the nearest time difference of current time, also comprise: according to history information service time and the clustering algorithm of described article to be recommended, determine the history cluster service time set of each article to be recommended; Described history information service time according to described article to be recommended and current time, determine the history service time of each article to be recommended and the nearest time difference of current time, be specially: according to the history cluster service time set of each article to be recommended, determine the cluster centre time of history service time of each article to be recommended; According to cluster centre time and the current time of history service time of each article to be recommended, determine the history service time of each article to be recommended and the nearest time difference of current time.

Further, described according to the history service time of each article to be recommended and the nearest time difference of current time, determine article recommendation list, be specially: according to the described history service time of each article to be recommended and the nearest time difference of current time and default time coefficient, determine the time score of each article to be recommended; According to the time score of each article to be recommended, determine article recommendation list.

Further, the degree of correlation of article and each article to be recommended is used to determine initial article recommendation list according to the history of user to be recommended, can be specially: the degree of correlation using article and each article to be recommended according to the history of described user to be recommended, determine the similarity score of each article to be recommended; According to the similarity score of each article to be recommended, determine initial article recommendation list.

Present invention also offers a kind of article recommendation server, comprise: the first acquisition module, history for obtaining article to be recommended uses information, and the history of described article to be recommended uses information to comprise the history use location information of article to be recommended and/or history information service time of article to be recommended, second acquisition module, for obtaining the current information of the terminal of user to be recommended, the current information of the terminal of described user to be recommended comprises current location information and/or the current time of the terminal of user to be recommended, and when the history of described article to be recommended uses information to comprise the history use location information of article to be recommended, the current information of the terminal of described user to be recommended comprises the current location information of the terminal of user to be recommended, when the history of described article to be recommended uses information to comprise history information service time of article to be recommended, the current information of the terminal of described user to be recommended comprises current time, first determination module, for using the current information of the terminal of information and described user to be recommended according to the history of described article to be recommended, determines article recommendation list, sending module, for sending to the terminal of described user to be recommended by described article recommendation list.

Further, described server also comprises: the 3rd acquisition module, uses Item Information for the history obtaining user to be recommended; Second determination module, for determining that the history of user to be recommended uses the degree of correlation of article and each article to be recommended; 3rd determination module, for using the degree of correlation of article and each article to be recommended according to the history of user to be recommended, determines initial article recommendation list; Described first determination module is further used for the current information using the terminal of information and user to be recommended according to the history of described initial article recommendation list, described each article to be recommended, determines article recommendation list.

Further, described first acquisition module is for obtaining the history use location information of article to be recommended; Described second acquisition module is for obtaining the current location information of the terminal of user to be recommended; Described first determination module comprises: the first determining unit, for the current location information according to the history use location information of described article to be recommended and the terminal of user to be recommended, determine the minimum distance of the current location of the history use location of each article to be recommended and the terminal of user to be recommended; Second determining unit, for the minimum distance of the current location according to the described history use location of each article to be recommended and the terminal of user to be recommended, determines article recommendation list.

Further, described first acquisition module is for obtaining history information service time of article to be recommended; Described second acquisition module is for obtaining current time; Described first determination module comprises: the 3rd determining unit, for according to history information service time of described article to be recommended and current time, determines the history service time of each article to be recommended and the nearest time difference of current time; 4th determining unit, for according to the described history service time of each article to be recommended and the nearest time difference of current time, determines article recommendation list.

Further, described first acquisition module is for history information service time of the history use location information and article to be recommended that obtain article to be recommended; Described second acquisition module is for obtaining current location information and the current time of the terminal of user to be recommended; Described first determination module comprises: the 5th determining unit, for the current location information according to the history use location information of described article to be recommended and the terminal of user to be recommended, determine the minimum distance of the current location of the history use location of each article to be recommended and the terminal of user to be recommended; 6th determining unit, for according to history information service time of described article to be recommended and current time, determines the history service time of each article to be recommended and the nearest time difference of current time; 7th determining unit, for the minimum distance of the current location according to the described history use location of each article to be recommended and the terminal of user to be recommended and the history service time of each article to be recommended and the nearest time difference of current time, determine article recommendation list.

Further, described server also comprises: the 4th determination module, for according to the history use location information of described article to be recommended and clustering algorithm, determines the history use location cluster set of each article to be recommended; Described first determining unit comprises: first determines subelement, for the history use location cluster set according to each article to be recommended, determines the cluster centre position of the history use location of each article to be recommended; Second determines subelement, and the current location for the cluster centre position of the history use location according to each article and the terminal of user to be recommended determines the minimum distance of the current location of the history use location of each article to be recommended and the terminal of user to be recommended.

Further, described second determining unit comprises: the 3rd determines subelement, for minimum distance and the default distance coefficient of the current location according to the described history use location of each article to be recommended and the terminal of user to be recommended, determine the distance score of each article to be recommended; 4th determines subelement, for the distance score according to each article to be recommended, determines article recommendation list.

Further, described server also comprises: the 5th determination module, for according to history information service time of described article to be recommended and clustering algorithm, determines the history cluster service time set of each article to be recommended; Described 3rd determining unit comprises: the 5th determines subelement, for the history cluster service time set according to each article to be recommended, determines the cluster centre time of history service time of each article to be recommended; 6th determines subelement, for cluster centre time and the current time of the history service time according to each article to be recommended, determines the history service time of each article to be recommended and the nearest time difference of current time.

Further, described 4th determining unit comprises: the 7th determines subelement, for according to the described history service time of each article to be recommended and the nearest time difference of current time and default time coefficient, determines the time score of each article to be recommended; 8th determines subelement, for the time score according to each article to be recommended, determines article recommendation list.

Further, described 3rd determination module comprises: the 8th determining unit, for using the degree of correlation of article and each article to be recommended according to the history of described user to be recommended, determines the similarity score of each article to be recommended; 9th determining unit, for the similarity score according to each article to be recommended, determines initial article recommendation list.

The beneficial effect of technique scheme of the present invention is as follows:

The invention provides a kind of item recommendation method and server, by obtaining the history use location of article, making do not have the article of position attribution also can participate in recommending, by obtaining the history service time of article, can recommend article based on service time.

Accompanying drawing explanation

The process flow diagram of the item recommendation method that Fig. 1 provides for the invention process 1.

The structural representation of the article recommendation server that Fig. 2 provides for the invention process 1.

Embodiment

For making the technical problem to be solved in the present invention, technical scheme and advantage clearly, be described in detail below in conjunction with the accompanying drawings and the specific embodiments.

The present invention is directed to the problem cannot recommended not having the article of position attribution and cannot recommend article based on service time in prior art, provide a kind of item recommendation method and server, can be recommended not having the article of position attribution by the history use location obtaining article, by obtaining the history service time of article, can recommend article based on service time.

The process flow diagram of the item recommendation method that Fig. 1 provides for the invention process 1, as shown in the figure, comprising:

S100, the history obtaining article to be recommended uses information, and the history of described article to be recommended uses information to comprise the history use location information of article to be recommended and/or history information service time of article to be recommended;

S102, obtain the current information of the terminal of user to be recommended, the current information of the terminal of described user to be recommended comprises current location information and/or the current time of the terminal of user to be recommended, and when the history of described article to be recommended uses information to comprise the history use location information of article to be recommended, the current information of the terminal of described user to be recommended comprises the current location information of the terminal of user to be recommended, when the history of described article to be recommended uses information to comprise history information service time of article to be recommended, the current information of the terminal of described user to be recommended comprises current time,

S104, uses the current information of the terminal of information and described user to be recommended, determines article recommendation list according to the history of described article to be recommended;

S106, sends to the terminal of described user to be recommended by described article recommendation list.

In technique scheme, the current information of the terminal of information and user to be recommended is used by the history obtaining article to be recommended, the current information of the terminal of information and user to be recommended is used according to the history of article to be recommended, determine article recommendation list, send it to the terminal of user to be recommended.Wherein, the history of article to be recommended uses information to comprise the history use location information of article to be recommended and/or history information service time of article to be recommended, and the current information of the terminal of user to be recommended comprises current location information and/or the current time of the terminal of user to be recommended.Thus, possess position attribution information without the need to article, use information can recommend article according to the history of article.

Each user to be recommended may have the preference of oneself to different article to be recommended, if do not consider that the recommendation list of the preference acquisition of user may can not meet the requirement of user.

In order to solve the problems of the technologies described above, further, the described history according to described article to be recommended uses the current information of the terminal of information and described user to be recommended, before determining article recommendation list, also comprises: the history obtaining user to be recommended uses Item Information; Determine that the history of user to be recommended uses the degree of correlation of article and each article to be recommended; Use the degree of correlation of article and each article to be recommended according to the history of user to be recommended, determine initial article recommendation list; The described history according to described each article to be recommended uses the current information of the terminal of information and described user to be recommended, determine article recommendation list, be specially: the current information using the terminal of information and described user to be recommended according to the history of described initial article recommendation list, described each article to be recommended, determine article recommendation list.

In technique scheme, the information of article is used by the history obtaining user to be recommended, namely the information of user to be recommended past used article is obtained, the degree of correlation of article and each article to be recommended is used to determine initial article recommendation list according to the history of user to be recommended, the degree of correlation of article such as can be determined according to the classification of article, when used article belong to same classification in the past for certain article to be recommended and user to be recommended, think that these article to be recommended are the article of user preference to be recommended, article are recommended to be placed in the prostatitis of initial article recommendation list on this band, also the degree of correlation of article can be determined according to the identical number of words of the key word of Item Title.The current information of the terminal of information and described user to be recommended is used according to the history of initial article recommendation list, described each article to be recommended, determine article recommendation list, thus, make article recommendation list consider the preference of user to be recommended, more can meet the requirement of user to be recommended.

The use location of some article has certain regularity usually; the use location of such as certain article only concentrates on specific place; if do not consider the use location of these article when recommending these article, recommendation results may be caused not meet the requirement of user.

In order to solve the problems of the technologies described above, further, the history of described article to be recommended uses information to comprise the history use location information of article to be recommended, the current information of the terminal of described user to be recommended comprises the current location information of the terminal of user to be recommended, the described history according to described article to be recommended uses the current information of the terminal of information and described user to be recommended, determine article recommendation list, be specially: according to the current location information of the history use location information of described article to be recommended and the terminal of user to be recommended, determine the minimum distance of the current location of the history use location of each article to be recommended and the terminal of user to be recommended, according to the minimum distance of the current location of the described history use location of each article to be recommended and the terminal of user to be recommended, determine article recommendation list.

In technique scheme, by the current location information of the terminal of the history use location information and user to be recommended that obtain article to be recommended, according to the minimum distance of the current location of the history use location information of each article to be recommended and the terminal of user to be recommended, determine article recommendation list.Thus the history use location made it possible to based on article is recommended article.Such as, there is article A to be recommended and article B, consider the history use location of article A and article B, the history use location of article A is place L1, and the history use location of article B is place L2, the current location of the terminal of user to be recommended is place L3, and wherein the distance of L1 and L3 is less than the distance of L2 and L3, then the final recommendation rank of article A will prior to the recommendation rank of article B.

The history use location information of described acquisition article to be recommended, can be specially: use record according to the article prestored, and create the use location list of article history, described article history use location list comprises the history use location of article; The described current location information according to the history use location information of described article to be recommended and the terminal of user to be recommended, determine the minimum distance of the current location of the history use location of each article to be recommended and the terminal of user to be recommended, can be specially: according to described article history use location list and the current location information of terminal recommending user, determine the minimum distance of the current location of the history use location of each article to be recommended and the terminal of user to be recommended.

In technique scheme, record is used to create the use location list of article history according to the article prestored, article use the history use location information including article in record, the article history use location information creating article history use location list in record is used according to article, thus, obtain the history use location of article.

The service time of some article may have certain regularity, such as, between 8 o'clock to 9 o'clock that only concentrate on service time of certain article the morning, if do not consider the service time of these article when recommending these article, recommendation results may be caused not meet the requirement of user.

In order to solve the problems of the technologies described above, further, the history of described article to be recommended uses information to comprise history information service time of article to be recommended, the current information of the terminal of described user to be recommended comprises current time, the described history according to described article to be recommended uses the current information of the terminal of information and described user to be recommended, determine article recommendation list, be specially: according to history information service time and the current time of described article to be recommended, determine the history service time of each article to be recommended and the nearest time difference of current time; According to the described history service time of each article to be recommended and the nearest time difference of current time, determine article recommendation list.

In technique scheme, by obtaining history information service time and the current time of article to be recommended, according to the history service time of article to be recommended and the nearest time difference of current time, determine article recommendation list.Thus, the history made it possible to based on article is recommended article service time, such as, there is article A to be recommended and article B, consider history service time of article A and article B, history service time of article A is 8 a.m. to 9 point, and history service time of article B is 2 pm to 3 point, current time is point in the morning 9, then the final recommendation rank of article A will prior to the recommendation rank of article B.

In technique scheme, by obtaining the history use location information of article to be recommended, history information service time, the current information of terminal of user to be recommended and current time, according to the described history use location of each article to be recommended and the minimum distance of current location of the terminal of user to be recommended and the nearest time difference of the history service time of article to be recommended and current time, determine article recommendation list.Thus, make it possible to recommend article service time based on the history use location information of article and history.

Because each article to be recommended may have multiple history use location, when the history use location of each article to be recommended is too much, during the minimum distance of the current location of the terminal in the history use location and user to be recommended of determining article to be recommended, the too much time can be expended.

In order to solve the problems of the technologies described above, further, the described current location information according to the history use location information of described article to be recommended and the terminal of user to be recommended, before determining the minimum distance of the current location of the history use location of each article to be recommended and the terminal of user to be recommended, also comprise: according to history use location information and the clustering algorithm of described article to be recommended, determine the history use location cluster set of each article to be recommended; The described current location information according to the history use location information of described article to be recommended and the terminal of user to be recommended, determine the minimum distance of the current location of the history use location of each article to be recommended and the terminal of user to be recommended, be specially: according to the history use location cluster set of each article to be recommended, determine the cluster centre position of the history use location of each article to be recommended; The minimum distance of the current location of the history use location of each article to be recommended and the terminal of user to be recommended is determined according to the current location of the cluster centre position of the history use location of each article to be recommended and the terminal of user to be recommended.

In technique scheme, according to clustering algorithm, the history use location of each article to be recommended is analyzed, obtain the history use location cluster set of each article to be recommended.Such as use K-means algorithm from the history use location information of each article to be recommended, first select arbitrarily k history use location information as initial cluster center; And for other history use location remaining, then according to the similarity (distance) of they and these cluster centres, respectively they are distributed to nearest with it (representated by cluster centre) cluster; And then calculate the cluster centre average of all positions (in this cluster) of each obtained new cluster; Constantly repeat this process until canonical measure function starts convergence.After the history use location cluster set obtaining each article to be recommended, determine the cluster centre position of the history use location of each article to be recommended, obtain the distance of the current location of the cluster centre position of the history use location of each article to be recommended and the terminal of user to be recommended, thus, the minimum distance of the current location of the history use location of each article to be recommended and the terminal of user to be recommended is determined with higher performance.

In technique scheme, each article to be recommended have a distance score, according to minimum distance and the default distance coefficient of the current location of the described history use location of each article to be recommended and the terminal of user to be recommended, determine the distance score of each article to be recommended; According to the distance score of each article to be recommended, determine article recommendation list.Such as, there is article A to be recommended, recommend article B, recommend article C and article D to be recommended, the distance of the current location of the history use location of article A to be recommended and the terminal of user to be recommended is 1km, the distance of the current location of the history use location of article B to be recommended and the terminal of user to be recommended is 2km, the distance of the current location of the history use location of article C to be recommended and the terminal of user to be recommended is 1.5km, the distance of the current location of the history use location of article D to be recommended and the terminal of user to be recommended is 0.5km, distance coefficient is 1, then the distance of article A to be recommended must be divided into 1, the distance of article B to be recommended must be divided into 2, the distance of article C to be recommended must be divided into 1.5, the distance of article D to be recommended must be divided into 0.5, article recommendation list is D, A, C, B.

Because each article to be recommended may have multiple history service time, when history service time of each article to be recommended is too much, when determining the nearest time difference of history service time of article to be recommended and current time, the too much time can be expended.

In order to solve the problems of the technologies described above, further, described history information service time according to described article to be recommended and current time, before determining the history service time of each article to be recommended and the nearest time difference of current time, also comprise: according to history information service time and the clustering algorithm of described article to be recommended, determine the history cluster service time set of each article to be recommended; Described history information service time according to described article to be recommended and current time, determine the history service time of each article to be recommended and the nearest time difference of current time, be specially: according to the history cluster service time set of each article to be recommended, determine the cluster centre time of history service time of each article to be recommended; According to cluster centre time and the current time of history service time of each article to be recommended, determine the history service time of each article to be recommended and the nearest time difference of current time.

In technique scheme, according to clustering algorithm, the history of each article to be recommended is analyzed service time, obtain the history cluster service time set of each article to be recommended.The above-mentioned method obtaining the history use location cluster set of each article to be recommended can be adopted to obtain the history cluster service time set of each article to be recommended.After the history cluster service time set obtaining each article to be recommended, determine the cluster centre time of history service time of each article to be recommended, obtain the cluster centre time of history service time of each article to be recommended and the difference of current time, thus, determine the history service time of each article to be recommended and the nearest time difference of current time with higher performance.

In technique scheme, each article to be recommended have a time score, according to the described history service time of each article to be recommended and the nearest time difference of current time and default time coefficient, determine the time score of each article to be recommended; According to the time score of each article to be recommended, determine article recommendation list.Such as, there is article A to be recommended, recommend article B, recommend article C and article D to be recommended, the history service time of article A to be recommended and the difference of current time are 1h, the history service time of article B to be recommended and the difference of current time are 2h, the history service time of article C to be recommended and the difference of current time are 1.5h, the history service time of article D to be recommended and the difference of current time are 0.5h, time coefficient is 1, then the time of article A to be recommended must be divided into 1, the time of article B to be recommended must be divided into 2, the time of article C to be recommended must be divided into 1.5, the time of article D to be recommended must be divided into 0.5, article recommendation list is D, A, C, B.

In technique scheme, each article to be recommended have a similarity score, such as, there is article A to be recommended, recommend article B, recommend article C and article D to be recommended, the history of user is recommended to use article to be article E and article F, the degree of correlation of article A to be recommended and article E is 1.5, is 0.5 with the degree of correlation of article F, then think that the history of article A to be recommended and user to be recommended uses the degree of correlation of article to be 1.5; The degree of correlation of article B to be recommended and article E is 1, is 2 with the degree of correlation of article F, then think that the history of article B to be recommended and user to be recommended uses the degree of correlation of article to be 2; The degree of correlation of article C to be recommended and article E is 0, is 0.5 with the degree of correlation of article F, then think that the history of article C to be recommended and user to be recommended uses the degree of correlation of article to be 0.5; The degree of correlation of article D to be recommended and article E is 1, is 2.5 with the degree of correlation of article F, then think that the history of article D to be recommended and user to be recommended uses the degree of correlation of article to be 2.5; Initial article recommendation list is D, B, A, C.

The structural representation of the article recommendation server that Fig. 2 provides for the invention process 1.As shown in the figure, described article recommendation server 20 comprises:

First acquisition module 21, uses information for the history obtaining article to be recommended, and the history of described article to be recommended uses information to comprise the history use location information of article to be recommended and/or history information service time of article to be recommended;

Second acquisition module 22, for obtaining the current information of the terminal of user to be recommended, the current information of the terminal of described user to be recommended comprises current location information and/or the current time of the terminal of user to be recommended, and when the history of described article to be recommended uses information to comprise the history use location information of article to be recommended, the current information of the terminal of described user to be recommended comprises the current location information of the terminal of user to be recommended, when the history of described article to be recommended uses information to comprise history information service time of article to be recommended, the current information of the terminal of described user to be recommended comprises current time,

First determination module 23, for using the current information of the terminal of information and described user to be recommended according to the history of described article to be recommended, determines article recommendation list;

Sending module 24, for sending to the terminal of described user to be recommended by described article recommendation list.

In technique scheme, the current information of the terminal of the user to be recommended that the history use information of the article each to be recommended that the first determination module 23 obtains according to the first acquisition module 21 and the second acquisition module 22 obtain, determine article recommendation list, article recommendation list is sent to the terminal of user to be recommended by sending module 24.Wherein, the history of article to be recommended uses information to comprise the history use location information of article to be recommended and/or history information service time of article to be recommended, and the current information of the terminal of user to be recommended comprises current location information and/or the current time of the terminal of user to be recommended.Thus, possess position attribution information without the need to article, use information can recommend article according to the history of article.

In the embodiment of the present invention 2, article use and are recorded as:

Employ article A at a place, position during at 9 in user X morning, during at 11 in the morning, employ article B at b place, position;

Employ article D at d place, position during at 9 in user Y morning, employ article A at e place, position during at 10 in the morning, during at 11 in the morning, employ article C at f place, position, during 2 pm, employ article E at c place, position;

Employ article B at a place, position during at 3 in user Z afternoon, employ article A at c place, position during at 5 in afternoon, during at 7 in the morning, employ article D at g place, position, during 8 a.m., employ article E at b place, position.

Then article to be recommended are article A, article B, article C, article D and article E.

For user to be recommended for X, user X carried terminal is in h place, position, and current time is point in the morning 10.

The history of user X uses article to be article A and article B, then the history of article A to be recommended and user X uses article similarity score to be 5, the history of article B to be recommended and user X uses article similarity score to be 5, the history of article C to be recommended and user X uses article similarity score to be 3.5, the history of article D to be recommended and user X uses article similarity score to be 3, and the history of article E to be recommended and user X uses article similarity score to be 4.In initial recommendation list be: A, B, E, C, D.Because article A and article B is the used article of user X, namely user X has known the relevant information of article A and article B, can not recommend to user X, and in order to make recommendation process simple, to user X used article no longer recommend, represent with the form of " article (score) ", namely obtaining initial recommendation list is E(4), C(3.5), D(3).

The distance coefficient preset is 1, and the history use location obtaining article C to be recommended from article use record is: position f; The history use location of article D to be recommended is: position d, position g; The history use location of article E to be recommended is: position b, position c; The distance of the current location h of position f and user X is 1km, then the minimum distance of the history use location of article C to be recommended and the current location of user X is 1km, and the distance of article C to be recommended must be divided into 1, and PTS and similarity score subtract distance and must be divided into: 2.5; The distance of position d and position h is 2km, and the distance of position g and position h is 1.5km, then the minimum distance of the history use location of article D to be recommended and the current location of user X is 1.5km, and the distance of article D to be recommended must be divided into 1.5, must be divided into 1.5; Position b be 3km with the distance of position h, the distance of position c and position h is 2km, then the minimum distance of the history use location of article E to be recommended and the current location of user X is 2km, and the distance of article E to be recommended must be divided into 2, must be divided into 2; Then obtaining article recommendation list is: C(2.5), E(2), D(1.5).

The time coefficient preset is 1, uses the history obtaining article C to be recommended record to be service time from article: the morning 11 point; The history of article D to be recommended is service time: the morning 7 point, the morning 9 point; The history of article E to be recommended is service time: 8 a.m., 2 pm; Because current time is point in the morning 10, then the history service time of article C to be recommended and the difference of current time are 1h, and the time of article C to be recommended must be divided into 1, and PTS and similarity score subtract distance score and subtract the time again and must be divided into: 1.5; The history service time of article D to be recommended and the difference of current time are 1h, and the time of article D to be recommended must be divided into 1, must be divided into 0.5; The history service time of article E to be recommended and the difference of current time are 2h, and the time of article E to be recommended must be divided into 2, must be divided into 0; Then obtaining article recommendation list is: C(1.5), D(0.5), E(0).

Namely finished article recommendation list is: C, D, E.

The above is the preferred embodiment of the present invention; it should be pointed out that for those skilled in the art, under the prerequisite not departing from principle of the present invention; can also make some improvements and modifications, these improvements and modifications also should be considered as protection scope of the present invention.

Claims

1. an item recommendation method, is characterized in that, comprising:

The history obtaining article to be recommended uses information, and the history of described article to be recommended uses information to comprise the history use location information of article to be recommended and/or history information service time of article to be recommended;

Obtain the current information of the terminal of user to be recommended, the current information of the terminal of described user to be recommended comprises current location information and/or the current time of the terminal of user to be recommended, and when the history of described article to be recommended uses information to comprise the history use location information of article to be recommended, the current information of the terminal of described user to be recommended comprises the current location information of the terminal of user to be recommended, when the history of described article to be recommended uses information to comprise history information service time of article to be recommended, the current information of the terminal of described user to be recommended comprises current time,

Use the current information of the terminal of information and described user to be recommended according to the history of described article to be recommended, determine article recommendation list;

Described article recommendation list is sent to the terminal of described user to be recommended.

2. method as claimed in claim 1, is characterized in that, the described history according to described article to be recommended uses the current information of the terminal of information and described user to be recommended, before determining article recommendation list, also comprises:

The history obtaining user to be recommended uses Item Information;

Determine that the history of user to be recommended uses the degree of correlation of article and each article to be recommended;

Use the degree of correlation of article and each article to be recommended according to the history of user to be recommended, determine initial article recommendation list;

The described history according to described article to be recommended uses the current information of the terminal of information and described user to be recommended, determines article recommendation list, is specially:

Use the current information of the terminal of information and described user to be recommended according to the history of described initial article recommendation list, described article to be recommended, determine article recommendation list.

3. method as claimed in claim 1, is characterized in that, the history of described article to be recommended uses information to comprise the history use location information of article to be recommended, and the current information of the terminal of described user to be recommended comprises the current location information of the terminal of user to be recommended,

According to the current location information of the history use location information of described article to be recommended and the terminal of user to be recommended, determine the minimum distance of the current location of the history use location of each article to be recommended and the terminal of user to be recommended;

According to the minimum distance of the current location of the described history use location of each article to be recommended and the terminal of user to be recommended, determine article recommendation list.

4. method as claimed in claim 1, is characterized in that, the history of described article to be recommended uses information to comprise history information service time of article to be recommended, and the current information of the terminal of described user to be recommended comprises current time,

According to history information service time and the current time of described article to be recommended, determine the history service time of each article to be recommended and the nearest time difference of current time;

According to the described history service time of each article to be recommended and the nearest time difference of current time, determine article recommendation list.

5. method as claimed in claim 1, it is characterized in that, the history of described article to be recommended uses information to comprise the history use location information of article to be recommended and history information service time of article to be recommended, the current information of the terminal of described user to be recommended comprises current location information and the current time of the terminal of user to be recommended

According to the minimum distance of the current location of the described history use location of each article to be recommended and the terminal of user to be recommended and the history service time of each article to be recommended and the nearest time difference of current time, determine article recommendation list.

6. method as described in claim 3 or 5, it is characterized in that, the described current location information according to the history use location information of described article to be recommended and the terminal of user to be recommended, before determining the minimum distance of the current location of the history use location of each article to be recommended and the terminal of user to be recommended, also comprise:

According to history use location information and the clustering algorithm of described article to be recommended, determine the history use location cluster set of each article to be recommended;

The described current location information according to the history use location information of described article to be recommended and the terminal of user to be recommended, determine the minimum distance of the current location of the history use location of each article to be recommended and the terminal of user to be recommended, be specially:

According to the history use location cluster set of each article to be recommended, determine the cluster centre position of the history use location of each article to be recommended;

The minimum distance of the current location of the history use location of each article to be recommended and the terminal of user to be recommended is determined according to the current location of the cluster centre position of the history use location of each article to be recommended and the terminal of user to be recommended.

7. method as claimed in claim 3, is characterized in that, the minimum distance of the described current location according to the history use location of each article to be recommended and the terminal of user to be recommended, determines article recommendation list, is specially:

According to minimum distance and the default distance coefficient of the current location of the described history use location of each article to be recommended and the terminal of user to be recommended, determine the distance score of each article to be recommended;

According to the distance score of each article to be recommended, determine article recommendation list.

8. method as described in claim 4 or 5, is characterized in that, described history information service time according to described article to be recommended and current time, before determining the history service time of each article to be recommended and the nearest time difference of current time, also comprises:

According to history information service time and the clustering algorithm of described article to be recommended, determine the history cluster service time set of each article to be recommended;

Described history information service time according to described article to be recommended and current time, determine the history service time of each article to be recommended and the nearest time difference of current time, be specially:

According to the history cluster service time set of each article to be recommended, determine the cluster centre time of history service time of each article to be recommended;

According to cluster centre time and the current time of history service time of each article to be recommended, determine the history service time of each article to be recommended and the nearest time difference of current time.

9. method as claimed in claim 4, is characterized in that, described according to the history service time of each article to be recommended and the nearest time difference of current time, determines article recommendation list, is specially:

According to the described history service time of each article to be recommended and the nearest time difference of current time and default time coefficient, determine the time score of each article to be recommended;

According to the time score of each article to be recommended, determine article recommendation list.

10. method as claimed in claim 2, is characterized in that, uses the degree of correlation of article and each article to be recommended to determine initial article recommendation list, can be specially according to the history of user to be recommended:

Use the degree of correlation of article and each article to be recommended according to the history of described user to be recommended, determine the similarity score of each article to be recommended;

According to the similarity score of each article to be recommended, determine initial article recommendation list.

11. 1 kinds of article recommendation servers, is characterized in that, comprising:

First acquisition module, uses information for the history obtaining article to be recommended, and the history of described article to be recommended uses information to comprise the history use location information of article to be recommended and/or history information service time of article to be recommended;

Second acquisition module, for obtaining the current information of the terminal of user to be recommended, the current information of the terminal of described user to be recommended comprises current location information and/or the current time of the terminal of user to be recommended, and when the history of described article to be recommended uses information to comprise the history use location information of article to be recommended, the current information of the terminal of described user to be recommended comprises the current location information of the terminal of user to be recommended, when the history of described article to be recommended uses information to comprise history information service time of article to be recommended, the current information of the terminal of described user to be recommended comprises current time,

First determination module, for using the current information of the terminal of information and described user to be recommended according to the history of described article to be recommended, determines article recommendation list;

Sending module, for sending to the terminal of described user to be recommended by described article recommendation list.

12. servers as claimed in claim 11, is characterized in that, also comprise:

3rd acquisition module, uses Item Information for the history obtaining user to be recommended;

Second determination module, for determining that the history of user to be recommended uses the degree of correlation of article and each article to be recommended;

3rd determination module, for using the degree of correlation of article and each article to be recommended according to the history of user to be recommended, determines initial article recommendation list;

Described first determination module is further used for the current information using the terminal of information and user to be recommended according to the history of described initial article recommendation list, described article to be recommended, determines article recommendation list.

13. servers as claimed in claim 11, is characterized in that,

Described first acquisition module is for obtaining the history use location information of article to be recommended;

Described second acquisition module is for obtaining the current location information of the terminal of user to be recommended;

Described first determination module comprises:

First determining unit, for the current location information according to the history use location information of described article to be recommended and the terminal of user to be recommended, determines the minimum distance of the current location of the history use location of each article to be recommended and the terminal of user to be recommended;

Second determining unit, for the minimum distance of the current location according to the described history use location of each article to be recommended and the terminal of user to be recommended, determines article recommendation list.

14. servers as claimed in claim 11, is characterized in that,

Described first acquisition module is for obtaining history information service time of article to be recommended;

Described second acquisition module is for obtaining current time;

Described first determination module comprises:

3rd determining unit, for according to history information service time of described article to be recommended and current time, determines the history service time of each article to be recommended and the nearest time difference of current time;

4th determining unit, for according to the described history service time of each article to be recommended and the nearest time difference of current time, determines article recommendation list.

15. servers as claimed in claim 11, is characterized in that,

Described first acquisition module is for history information service time of the history use location information and article to be recommended that obtain article to be recommended;

Described second acquisition module is for obtaining current location information and the current time of the terminal of user to be recommended;

Described first determination module comprises:

5th determining unit, for the current location information according to the history use location information of described article to be recommended and the terminal of user to be recommended, determines the minimum distance of the current location of the history use location of each article to be recommended and the terminal of user to be recommended;

6th determining unit, for according to history information service time of described article to be recommended and current time, determines the history service time of each article to be recommended and the nearest time difference of current time;

7th determining unit, for the minimum distance of the current location according to the described history use location of each article to be recommended and the terminal of user to be recommended and the history service time of each article to be recommended and the nearest time difference of current time, determine article recommendation list.

16. servers as claimed in claim 13, is characterized in that, also comprise:

4th determination module, for according to the history use location information of described article to be recommended and clustering algorithm, determines the history use location cluster set of each article to be recommended;

Described first determining unit comprises:

First determines subelement, for the history use location cluster set according to each article to be recommended, determines the cluster centre position of the history use location of each article to be recommended;

Second determines subelement, and the current location for the cluster centre position of the history use location according to each article and the terminal of user to be recommended determines the minimum distance of the current location of the history use location of each article to be recommended and the terminal of user to be recommended.

17. servers as claimed in claim 13, it is characterized in that, described second determining unit comprises:

3rd determines subelement, for minimum distance and the default distance coefficient of the current location according to the described history use location of each article to be recommended and the terminal of user to be recommended, determines the distance score of each article to be recommended;

4th determines subelement, for the distance score according to each article to be recommended, determines article recommendation list.

18. servers as claimed in claim 14, is characterized in that, also comprise:

5th determination module, for according to history information service time of described article to be recommended and clustering algorithm, determines the history cluster service time set of each article to be recommended;

Described 3rd determining unit comprises:

5th determines subelement, for the history cluster service time set according to each article to be recommended, determines the cluster centre time of history service time of each article to be recommended;

6th determines subelement, for cluster centre time and the current time of the history service time according to each article to be recommended, determines the history service time of each article to be recommended and the nearest time difference of current time.

19. servers as claimed in claim 14, it is characterized in that, described 4th determining unit comprises:

7th determines subelement, for according to the described history service time of each article to be recommended and the nearest time difference of current time and default time coefficient, determines the time score of each article to be recommended;

8th determines subelement, for the time score according to each article to be recommended, determines article recommendation list.

20. servers as claimed in claim 12, it is characterized in that, described 3rd determination module comprises:

8th determining unit, for using the degree of correlation of article and each article to be recommended according to the history of described user to be recommended, determines the similarity score of each article to be recommended;

9th determining unit, for the similarity score according to each article to be recommended, determines initial article recommendation list.