JP2010003134A - Server, method, and program for recommending retrieval keyword - Google Patents

Abstract

A server, method, and program capable of recommending useful keywords belonging to the same field at the time of search are provided.
A server acquires a phrase used by each of a plurality of users and the number of times the phrase is used, and stores in advance a user attribute DB that stores the phrase in association with the user, a category, and a category determination keyword that characterizes the category. Category reader DB 22 to be used, and a category leader extraction unit that determines a user belonging to the category as a representative user based on the similarity between the stored phrase and the category determination keyword and the number of times the phrase having a high similarity is used 13 and according to the input of the search keyword, the search keyword is classified based on the category determination keyword, representative users belonging to the classified category are extracted, and stored in association with the extracted representative user. Search control that extracts selected phrases as recommended keywords It includes a 14, a.
[Selection] Figure 3

Description

  The present invention relates to a server, a method, and a program for presenting a new recommended keyword for an input search keyword.

  2. Description of the Related Art Conventionally, a web page search system on the Internet receives a keyword input and extracts a web page including the keyword and provides it to the user. In such a system, since the search result depends on the keyword input by the user, the same Web page is not always searched even for similar keywords.

Therefore, a technique for adding or replacing a recommended keyword has been proposed in order to make it easy to search for a Web page desired by the user. For example, Patent Document 1 shows that keywords having a co-occurrence relationship are grouped from a keyword history related to content used by a user, and keywords in the same group are added or replaced. .
JP 2005-310094 A

  However, in the method of Patent Document 1, since the recommended keywords are limited to keywords having a co-occurrence relationship, it is not possible to recommend useful keywords that belong to the same field but have no co-occurrence relationship.

  By the way, among users who are interested in the field related to the search keyword, there are those who use special keywords such as a term unique to the field, the latest term, or a name that is not widely known. In many cases, a Web page related to such a special keyword is required.

  Accordingly, an object of the present invention is to provide a server, a method, and a program capable of recommending useful keywords belonging to the same field (category) at the time of search.

  The present invention provides the following solutions.

(1) A server that presents a new recommended keyword for an input search keyword,
As attribute of each of a plurality of users, the attribute storage means for acquiring the word used by each user and the number of times the word is used, and storing it in association with the user;
Category storage means for storing in advance a category and at least one category determination keyword characterizing the category;
For all of the category determination keywords stored by the category storage unit, calculate the similarity to the phrase stored by the attribute storage unit and the number of times the phrase is used, and calculate the similarity and the A determination means for determining the user belonging to the category as a representative user based on the number of times the phrase having a high similarity is used;
Determination means for classifying the search keyword based on the category determination keyword in response to the input of the search keyword, and determining a category of the search keyword;
A server comprising: extraction means for extracting the representative user belonging to the category determined by the determination means, and extracting the phrase stored in the attribute storage means in association with the extracted representative user as the recommended keyword .

  According to such a configuration, the server acquires the phrase used by each user and the number of times the phrase is used as an attribute of each of the plurality of users, stores it in association with the user, and characterizes the category and the category. At least one category determination keyword is stored in advance, and for all of the category determination keywords, the similarity to the stored word and the number of times of use of the word having a high similarity are calculated, and the calculated similarity and the similarity A user belonging to the category is determined as a representative user based on the number of times the phrase is used, and according to the input of the search keyword, the search keyword is classified based on the category determination keyword, Judge the category of the search keyword and extract representative users belonging to the judged category The phrase stored in association with the representative user that the extracted is extracted as recommended keyword.

  As a result, the server determines a representative user who uses more frequently and more frequently words / phrases with high similarity to the category determination keywords for each category, and extracts the words / phrases used by the representative user as recommended keywords. There is a possibility that useful keywords belonging to the same category can be recommended for the search keyword input by the user.

  Furthermore, since the server determines the representative user based on the similarity to the category determination keyword, it is possible to determine a key person who is familiar with the category. Therefore, a general user can use a keyword with high freshness used by a key person who leads a category for a search.

  (2) The server according to (1), wherein the attribute storage unit stores words / phrases used in a predetermined period among words / phrases used by the user.

  According to such a configuration, since the server stores words / phrases used within a predetermined period, for example, within one month or within one year, it eliminates words / phrases that are no longer used with the passage of time. Can be reflected. Thus, there is a possibility that an appropriate representative user can be determined for each category.

  (3) The server according to (1) or (2), wherein the determination unit determines a new representative user at a predetermined cycle.

  According to such a configuration, since the server determines a new representative user at a predetermined cycle, an appropriate representative user can be determined and updated as time passes. As a result, there is a possibility that an appropriate keyword can be recommended to the user who has input the search keyword.

(4) The determination unit ranks and determines a plurality of representative users based on the similarity,
The server according to any one of (1) to (3), wherein the extracting unit assigns a priority to the recommended keyword based on the ranking of the representative user by the determining unit.

  According to such a configuration, the server ranks and determines a plurality of representative users based on the similarity between the stored phrase and the category determination keyword, and recommends based on the ranking of the representative users. Give priority to keywords.

  As a result, the server gives priority to the recommended keywords based on the rankings for a plurality of representative users, so that it is possible to preferentially present useful recommended keywords to the user who has input the search keyword. There is sex.

  (5) Regarding the category, the classification rule for the category is determined based on the phrase used by the representative user determined by the determining unit, and the phrase used by a user different from the representative user is determined. The server according to any one of (1) to (4), further including an adding unit that adds a word / phrase classified into the category as the recommended keyword when classified into the category by a rule.

  According to such a configuration, the server learns a classification rule for a category based on a phrase used by a representative user, and calculates a degree of association with the category even for a phrase used by another user. be able to. Then, the server can present a word / phrase determined to be classified into this category as a recommended keyword with a high degree of association.

  Thus, the server can recommend words or phrases that are closely related to the category of the input search keyword to the user even if the words are used by a user other than the representative user.

(6) A method in which a computer presents a new recommended keyword for an input search keyword,
As an attribute of each of a plurality of users, an attribute storage step of acquiring a phrase used by each user and the number of times of use of the phrase, and storing it in association with the user;
A category storing step for storing in advance a category and at least one category determination keyword characterizing the category;
For all of the category determination keywords stored in the category storage step, calculate the similarity to the word stored in the attribute storage step and the number of uses of the word with high similarity, and calculate the similarity and the A determination step of determining, as a representative user, the user belonging to the category based on the number of times the phrase having a high similarity is used;
A determination step of classifying the search keyword based on the category determination keyword and determining a category of the search keyword in response to the input of the search keyword;
An extraction step of extracting the representative user belonging to the category determined in the determination step and extracting the phrase stored in the attribute storage step in association with the extracted representative user as the recommended keyword. .

  According to such a configuration, the same effect as in (1) can be expected by executing the method.

  (7) A program that causes a computer to execute the method according to (6).

  According to such a configuration, the same effect as in (1) can be expected by causing the computer to execute the program.

  ADVANTAGE OF THE INVENTION According to this invention, when a user searches, the useful keyword which belongs to the same category can be recommended.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

[Entire system configuration]
FIG. 1 is a diagram illustrating an overall configuration of a server 10 and related elements according to the present embodiment. The server 10 is connected to the user terminals 20 and 30 via the network 40.

  The server 10 collects, from the user terminal 20, for each user, a search query input by each user and a phrase extracted from a web page such as a blog. And the server 10 collects the phrase used for every user, and the frequency | count of its use, compares each of several users, and determines the representative user (category leader) who represents a category.

  Thereafter, when a search keyword is input from the user terminal 30, the server 10 presents another recommended keyword belonging to the same category as the search keyword based on the phrase used by the representative user.

[Hardware configuration]
FIG. 2 is a diagram illustrating a hardware configuration of the server 10 according to the present embodiment. The server 10 includes a CPU (Central Processing Unit) 310 (a plurality of CPUs such as a CPU 320 may be added in a multiprocessor configuration), a bus line 200, and a communication I / F (I / F: interface) that configure the control unit 300. 330, a main memory 340, a BIOS (Basic Input Output System) 350, an I / O controller 360, a hard disk 370, an optical disk drive 380, and a semiconductor memory 390. The hard disk 370, the optical disk drive 380, and the semiconductor memory 390 are collectively referred to as a storage device 410.

  The control unit 300 is a part that controls the server 10 in an integrated manner, and appropriately reads and executes various programs stored in the hard disk 370 (described later), thereby cooperating with the above-described hardware and according to the present invention. Various functions are realized.

  The communication I / F 330 is a network adapter when the server 10 transmits / receives information to / from other devices such as the user terminals 20 and 30 in FIG. 1 via the network. The communication I / F 330 may include a modem, a cable modem, and an Ethernet (registered trademark) adapter.

  The BIOS 350 records a boot program executed by the CPU 310 when the server 10 is started up, a program depending on the hardware of the server 10, and the like.

  A storage device 410 such as a hard disk 370, an optical disk drive 380, and a semiconductor memory 390 can be connected to the I / O controller 360.

  The hard disk 370 stores various programs for causing the hardware to function as the server 10, a program for executing the functions of the present invention, a table to be described later, and the like. The server 10 can also use an external hard disk (not shown) as an external storage device.

  As the optical disk drive 380, for example, a DVD-ROM drive, a CD-ROM drive, a DVD-RAM drive, or a CD-RAM drive can be used. In this case, the optical disk 400 corresponding to each drive is used. A program or data can be read from the optical disk 400 by the optical disk drive 380 and provided to the main memory 340 or the hard disk 370 via the I / O controller 360.

  Note that the computer in the present invention refers to an information processing device including a storage device, a control unit, and the like, and the server 10 includes an information processing device including a storage device 410, a control unit 300, and the like. The apparatus is included in the computer concept of the present invention.

[Function configuration]
FIG. 3 is a diagram illustrating a functional configuration of the server 10 according to the present embodiment. The control unit 300 of the server 10 includes a user attribute analysis unit 11, a category specification unit 12, a category reader extraction unit 13, and a search control unit 14. The storage device 410 of the server 10 includes a user attribute DB 21 and a category reader DB 22.

  The user attribute analysis unit 11 receives text in a Web page such as a blog created by a user, a search query input at the time of information search, and the like from the user terminal 20 and extracts words included in these. Then, the user attribute analysis unit 11 selects a predetermined type of keyword from these phrases, and stores this keyword and the number of times the keyword has been used as the user attribute in the user attribute DB 21.

  Here, the predetermined types of keywords may be set in advance as, for example, nouns or verbs. Moreover, it is good also as selecting an important word by calculating the importance based on the appearance frequency of a phrase.

  The user attribute analysis unit 11 selects a Web page created during a predetermined period such as one month or one year until now, or an input search query in order to reflect the recent usage trend of the keyword, The keywords used for a predetermined period may be stored. Moreover, since it is preferable that the stored keyword is always new, the user attribute analysis unit 11 repeats the processing at a predetermined cycle to keep the user attribute DB 21 in the latest state.

  FIG. 4 is a diagram showing a user attribute table storing keywords as user attributes according to the present embodiment. Here, for the user ID, the keywords extracted from the blog text and the keywords extracted from the search query are stored together with the number of times they are used.

  Here, a field is prepared for each keyword information source in the user attribute table. However, the field is not limited to this, and it may be stored together in a single field. In addition, according to storing by information source, the server 10 can classify and present recommended keywords by information source.

  The category specifying unit 12 receives a category specification from the administrator of the server 10. Specifically, for example, a plurality of categories such as “Akiba” and “Bike” are set, and input for specifying one or more keywords belonging to each category is accepted. Then, the received category designation data is stored in the category reader DB 22.

  Based on the category designation data stored by the category designation unit 12, the category leader extraction unit 13 extracts a category leader that is a user representing each category from the user attribute DB 21. There may be a plurality of category leaders to be extracted, and the category leader extraction unit 13 stores the category leader identification data in the category leader DB 22 in association with each category.

  Here, when extracting the category leader, the category leader extraction unit 13 blog text stored in the user attribute table (FIG. 4) for all of the category determination keywords stored in the category reader table (FIG. 5). The degree of coincidence with keywords such as or queries is calculated. Here, for each category determination keyword, as the number of similar keywords used increases, the degree of matching is weighted and the user attribute is highly evaluated. Then, the category leader extraction unit 13 extracts, as a category leader, a user whose evaluation value is higher than a predetermined number or an evaluation value equal to or higher than a predetermined value. Thereby, a user who uses more keywords similar to the category determination keywords and uses them more frequently is highly evaluated as a whole and is extracted as a category leader.

  In the above-described evaluation, how much weighting is given to each of the similarity calculated for each category determination keyword, the number of times of use, and the number of category determination keywords for which it is determined that similar words are stored in the user attribute table. It may be designed as appropriate while actually verifying the extracted category leader.

  Further, the extraction method is not limited to this. For example, the keywords in the user attribute table (FIG. 4) are classified based on the classification rule learned by the category determination keyword, and the users who have a strong similarity tendency to the category are classified as category leaders. May be extracted as

  FIG. 5 is a diagram showing a category reader table storing categories and category readers according to the present embodiment.

  Here, first, the category determination keyword is stored for the category ID and the category name by the category specification received by the category specifying unit 12. The category leader ID extracted by the category leader extraction unit 13 is stored in association with each category.

  Further, the category leader extraction unit 13 determines a classification rule for the category based on the category determination keyword and the keyword stored in the user attribute table (FIG. 4) corresponding to the category leader, and stores the classification rule in association with the category. May be. As a result, even keywords that are not associated with each other can be categorized, so there is a possibility that many recommended keywords can be presented.

  Note that existing techniques can be used to determine the classification rules, and can be realized by a learning machine such as SVM (Support Vector Machine), for example.

  Here, the category leader extraction unit 13 determines a new category leader at a predetermined cycle, and updates the category leader table (FIG. 5). Thus, as the user attribute table (FIG. 4) is updated, the category leader is also updated, so that the latest category leader can always be selected for each category.

  In the category leader table (FIG. 5), the category leader extraction unit 13 may rank the category leaders. As a result, the importance of the keyword used by the upper category leader can be determined to be high.

  The search control unit 14 extracts a recommended keyword based on the user attribute DB 21 and the category reader DB 22 and transmits it to the user terminal 30 in response to receiving the search keyword input in the user terminal 30.

  At this time, the search control unit 14 refers to the category leader table (FIG. 5) and determines the category to which the received search keyword belongs. Then, the search control unit 14 extracts a category leader of the determined category, and extracts a keyword used by the category leader stored in the user attribute table (FIG. 4) as a recommended keyword. Note that the recommended keywords may be extracted by ranking according to the number of uses.

  The category to which the search keyword belongs can be determined by detecting a match or similarity with the category determination keyword in the category leader table (FIG. 5). Alternatively, the search keyword may be classified into any category based on the classification rule determined for the category.

  Here, the search control unit 14 extracted the keyword used by the category reader as the recommended keyword, but is not limited thereto. For example, keywords used by users other than the category leader are also extracted from the user attribute table (FIG. 4) and classified based on the category classification rule. May be transmitted to the user terminal 30. Thereby, not only a category leader but a useful keyword can be recommended from the keywords used by various users.

  The search control unit 14 weights the keywords used by the category leader based on the ranking of the category leaders ranked by the category leader extraction unit 13. This gives priority to the recommended keywords transmitted to the user terminal 30, so that there is a possibility that useful keywords can be preferentially recommended to the user.

[Processing flow]
FIG. 6 is a flowchart showing a category leader determination process in the server 10 according to the present embodiment.

  In step S1, the control unit 300 accepts a category designation input from the administrator of the server 10 and stores it in the category reader table (FIG. 5).

  In step S2, the control unit 300 reads keywords (blog text, search query, etc.) used by each user and the number of times of use as user attributes from the user attribute table (FIG. 4).

  In step S3, the control unit 300 calculates the similarity between the category determination keyword in the category leader table (FIG. 5) and the user attribute read in step S2.

  In step S4, the control unit 300 selects one or more category leaders for each category based on the similarity calculated in step S3 and the number of times the keyword having a high similarity is used. The selected category leader is stored in the category leader table (FIG. 5).

  In step S5, the control unit 300 learns the classification rule for each category by a predetermined learning machine based on the keyword used by the category leader selected in step S4, and stores it in the category reader table (FIG. 5). To do.

  FIG. 7 is a flowchart showing recommended keyword output processing in the server 10 according to the present embodiment.

  In step S <b> 11, the control unit 300 receives a search keyword input at the user terminal 30.

  In step S12, the control unit 300 classifies the search keyword received in step S11 into one of the categories based on the category determination keyword in the category leader table (FIG. 5). In addition, when it is not similar to any category determination keyword, it does not classify into any category.

  In step S13, the control unit 300 determines whether or not the search keyword is classified into any category in step S12. If this determination is YES, the process proceeds to step S14, and if the determination is NO, the recommended keyword is not output and the process ends.

  In step S14, the control unit 300 extracts the category leader selected for the category classified in step S12, and extracts the keywords used by the category leader from the user attribute table (FIG. 4).

  In step S15, the control unit 300 determines whether or not the keywords used by the user different from the category leader extracted in step S14 are classified into the same category. Then, the control unit 300 extracts keywords determined to be classified.

  In step S16, the control unit 300 transmits the keyword extracted in steps S14 and S15 as a recommended keyword to the user terminal 30, and displays the screen as a list that can be selected by the user.

  In the present embodiment, the recommended keywords are extracted according to the input of the search keyword (steps S14 to S15), but the present invention is not limited to this. The controller 300 may extract and store recommended keywords for each category in advance. According to this, since the recommended keyword can be quickly transmitted to the user terminal 30, the operability for the user is improved.

  As mentioned above, although embodiment of this invention was described, this invention is not restricted to embodiment mentioned above. The effects described in the embodiments of the present invention are only the most preferable effects resulting from the present invention, and the effects of the present invention are limited to those described in the embodiments of the present invention. is not.

It is a figure which shows the whole structure of the server 10 and related element which concern on embodiment of this invention. It is a figure which shows the hardware constitutions of the server 10 which concerns on embodiment of this invention. It is a figure which shows the function structure of the server 10 which concerns on embodiment of this invention. It is a figure which shows the user attribute table which concerns on embodiment of this invention. It is a figure which shows the category leader table which concerns on embodiment of this invention. It is a flowchart which shows the determination process of the category leader which concerns on embodiment of this invention. It is a flowchart which shows the output process of the recommendation keyword which concerns on embodiment of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Server 11 User attribute analysis part 12 Category designation | designated part 13 Category reader extraction part 14 Search control part 21 User attribute DB
22 Category Reader DB
20, 30 User terminal 40 Network 300 Control unit 410 Storage device

Claims (7)

A server that presents a new recommended keyword for an input search keyword,
As attribute of each of a plurality of users, the attribute storage means for acquiring the word used by each user and the number of times the word is used, and storing it in association with the user;
Category storage means for storing in advance a category and at least one category determination keyword characterizing the category;
For all of the category determination keywords stored by the category storage unit, calculate the similarity to the phrase stored by the attribute storage unit and the number of times the phrase is used, and calculate the similarity and the A determination means for determining the user belonging to the category as a representative user based on the number of times the phrase having a high similarity is used;
Determination means for classifying the search keyword based on the category determination keyword in response to the input of the search keyword, and determining a category of the search keyword;
A server comprising: extraction means for extracting the representative user belonging to the category determined by the determination means, and extracting the phrase stored in the attribute storage means in association with the extracted representative user as the recommended keyword .   The server according to claim 1, wherein the attribute storage unit stores words used in a predetermined period among words used by the user.   The server according to claim 1, wherein the determination unit determines a new representative user at a predetermined cycle. The determining means ranks and determines a plurality of representative users based on the similarity,
4. The server according to claim 1, wherein the extraction unit assigns a priority to the recommended keyword based on ranking of representative users by the determination unit. 5.   Regarding the category, a classification rule for the category is determined based on the phrase used by the representative user determined by the determination unit, and a phrase used by a user different from the representative user is determined according to the determined rule. The server according to any one of claims 1 to 4, further comprising an adding unit that, when classified into a category, adds a phrase classified into the category as the recommended keyword. A method in which a computer presents a new recommended keyword for an input search keyword,
As an attribute of each of a plurality of users, an attribute storage step of acquiring a phrase used by each user and the number of times of use of the phrase, and storing it in association with the user;
A category storing step for storing in advance a category and at least one category determination keyword characterizing the category;
For all of the category determination keywords stored in the category storage step, calculate the similarity to the word stored in the attribute storage step and the number of uses of the word with high similarity, and calculate the similarity and the A determination step of determining, as a representative user, the user belonging to the category based on the number of times the phrase having a high similarity is used;
A determination step of classifying the search keyword based on the category determination keyword and determining a category of the search keyword in response to the input of the search keyword;
Extracting the representative user belonging to the category determined in the determining step, and extracting the phrase stored in the attribute storing step in association with the extracted representative user as the recommended keyword. .   The program which makes a computer perform the method of Claim 6.

Images