JP4509918B2 - Information retrieval method and apparatus, and computer-readable recording medium - Google Patents

Description

  The present invention relates to an information search method and apparatus and a computer-readable recording medium, and in particular, information search capable of managing user operation history information and easily searching for information browsed by the user in the past. The present invention relates to a method and apparatus and a computer-readable recording medium.

  Conventionally, as a method of searching for user operation history, a detailed operation history of the user terminal is collected, and an empirical score indicating the strength experienced by the user is calculated for each file browsed by the user in the past. In addition, there is a method of easily searching for an important file among files browsed by the user in the past (see, for example, Patent Document 1).

In addition, a tool for displaying a search result of a file browsed by a user in the past and a search result of a file on the Internet is disclosed.
JP 2005-174138 A “Content Search Method and System, Program, and Program Recording Medium”

  When a user wants to obtain “information about a certain matter” and searches using Google (registered trademark) or goo (registered trademark), there are hundreds or thousands of search results. Not all search results can be confirmed. Specifically, Google (registered trademark) and goo (registered trademark) display in the degree of suitability to the search keyword by a unique ranking method. Ranking is fixed.

  However, Web pages viewed in the past may be more effective for the user. For example, a Web page that has been viewed in the past is easier to understand “information about a certain matter” than a Web page that has not been viewed in the past. Since it is possible to guess to some extent what information is written, it is possible to reach “information about a certain matter” in the Web page in a short time, and to understand “information about a certain matter” by suppressing points. And the stronger the degree that the user has experienced in the past, the higher the possibility that the user can understand in a short time.

  And with existing technology, search from a page you've seen in the past, using a small amount of operation history, and display the search results, or search regardless of whether you've seen it in the past However, the search results are displayed without considering that the page has been viewed in the past. In other words, an experience score indicating the strength experienced by the user cannot be calculated sufficiently.

  Also, in the conventional method, since the search result of the file browsed by the user in the past and the search result of the file on the Internet are displayed separately, each file indicated by the search result on the Internet is displayed in the past. I don't know how related to the file I browsed.

  The present invention has been made in view of the above points, an information search method and apparatus, and a computer-readable record capable of easily searching for information that has been strongly experienced in the past that seems to be valuable for browsing again. To provide a medium.

  FIG. 1 is a diagram for explaining the principle of the present invention.

The present invention (Claim 1) is an information retrieval method for retrieving operation history information,
When a search request is received from the request source, a file search step (step 1) in which the file search means searches for an information body (file) browsed in the past by a user who matches the search request based on the search request. ,
Based on the storage location information of the file, the score calculation means indicates the strength that the user has experienced using the file suitability score indicating the degree of suitability for the search request for each file or the operation history information in the operation history database. A score calculation step (step 2) for calculating an experience score represented for each file operated on the PC, and calculating an integrated score obtained by integrating the file fitness score and the experience score;
Similarity calculation means for calculating a similarity between a file obtained as a result of searching an arbitrary range of information including files not browsed by the user and a file browsed in the past (step 3, 4)
The search management means performs an output step (steps 5 and 6) for outputting a search result based on the display priority obtained from the similarity and the integrated score.
And do.

The present invention (Claim 2) receives, from the requester, an identifier that specifies a method for calculating a file suitability score, an experience score, and an integrated score for a search request;
In the score calculation step,
According to the identifier, the method of calculating the file suitability score, the experience score, and the integrated score for the search request is changed, and the file suitability score, the experience score, and the integrated score are calculated.

Further, the present invention (Claim 3) receives, from the request source, display priority adjustment information that specifies a method for calculating the display priority from the similarity and the integrated score;
Calculating display priority using display priority adjustment information.

In the present invention (Claim 4), the operation history acquisition means includes:
Using a message message on a PC and communication with a program operating on the PC to obtain a history of operating the user's PC;
Obtaining an identifier for identifying information viewed by the user from a history of operating the user's PC;
Based on an identifier for identifying information browsed by the user, obtaining the information body browsed in the past by the user, and storing it in the operation history database as a file;
Acquiring attribute information of the information body and storing it in the file attribute information database as file attribute information.

  FIG. 2 is a principle configuration diagram of the present invention.

The present invention (Claim 5) is an information retrieval apparatus for retrieving operation history information,
The operation history database 109 includes operation history information which is a history of operating a user's personal computer (PC),
Based on a search request acquired from a request source, a file search means 103 for searching an information body (file) browsed in the past by a user who matches the search request;
Based on the storage location information of the file, the file conformity score indicating the degree of conformity to the search request for each file or the strength experienced by the user using the operation history information of the operation history database 109 is operated on the PC. A score calculation unit 112 that calculates an experience score represented for each file and calculates an integrated score obtained by integrating the file suitability score and the experience score;
A similarity calculation means 107 for obtaining a similarity between a file as a result of searching information in an arbitrary range including a file that is not browsed by a user and a file browsed in the past;
Output means 102 for outputting a search result based on the display priority obtained from the similarity and the integrated score.

The present invention (Claim 6) provides a computer with:
A computer-readable recording medium storing an information search program for executing each step of the information search method according to claim 1.

  As described above, according to the present invention, when searching for information on the Internet, the individual files shown in the search results are viewed in the past in consideration of the strength experienced in the past. A certain web page can be displayed. Files that have been experienced more strongly in the past are more likely to be representative files of actions that have been taken in the past. You can also remember. In addition, information that has been strongly experienced in the past can be acquired in a relatively short time when viewed again, compared to information that has been weakly experienced in the past (for example, only for a short time). According to the present invention, it is possible to easily search for information that has been strongly experienced in the past that is considered worthwhile to be browsed again.

  Further, it is possible to easily find a file browsed in the past similar to the contents of search results on the Internet and URLs.

  In addition, the user can easily specify which of the similarities between contents and URL and the strength experienced in the past is prioritized to display a file browsed in the past.

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

  FIG. 3 shows a configuration of the search device according to the embodiment of the present invention.

  A search device 100 shown in the figure is connected to an information search terminal (IE) 200, an external search engine (goo (registered trademark)) 300, and a Web server 400.

  Based on a search request from the information search terminal 200, the search device 100 adds and presents a file that has been viewed in the past to each file of the search result on the Internet.

  The search device 100 includes an operation history acquisition unit 101, a search management unit 102, a file search unit 103, an experience score calculation unit 104, an integrated score calculation unit 105, an all page search unit 106, a similarity calculation unit 107, and a display priority calculation unit. 108, an operation history DB (database) 109, a file attribute information DB (database) 110, and a past operation file storage unit 111.

  The operation history acquisition unit 101 uses a message message on a PC and communication with a program operating on the PC to acquire a history of operating the user's PC and operate the user's PC. An identifier for identifying information browsed by the user is acquired from the recorded history, and based on an identifier for identifying information browsed by the user, an information body browsed in the past by the user is acquired and operated as a file. Save in the history database 109. Further, the operation history acquisition unit 101 acquires attribute information of the information main body and stores it in the file attribute information DB 110 as file attribute information.

The search management unit 102 acquires the following information from the information search terminal 200.
・ Search request (search keyword);
An identifier that specifies how to calculate the file suitability score for a search request;
An identifier that specifies how to calculate the integrated score for the search request;
An identifier that specifies how to calculate the experience score for a search request;
Display priority adjustment information for specifying a method for calculating the display priority from the similarity and the integrated score;
The search request and an identifier for designating a method for calculating the file suitability score for the search request are transferred to the file search unit 103. In addition, an identifier that specifies a method for calculating the experience score for the search request is transferred to the experience score calculation unit 104. In addition, an identifier that specifies a method for calculating the integrated score for the search request is transferred to the integrated score calculation unit 105. Further, display priority adjustment information for specifying a method for calculating the display priority from the similarity and the integrated score is transferred to the display priority calculation unit 108. Further, the search request is transferred to the all page search unit 106.

  In addition, the search management unit 102 interacts with each component, and display priority is given to a result obtained by searching information in an arbitrary range including files that are not viewed by the user acquired from the total score calculation unit 105 as a population. A search result display file for associating and displaying a file browsed in the past with a high display priority acquired from the degree calculation unit 108 is created and transmitted to the information search terminal 200 of the search request source.

  Based on the search request from the request source, the file search unit 103 acquires storage location information of the information body (file) browsed in the past by the user from the operation history database 109, and based on the storage location information, the user The file browsed in the past is searched by the text and file attributes in the file of the past operation file storage unit 111. Further, based on an identifier that specifies a method for calculating the file suitability score for the search request acquired from the information search terminal 200, a file suitability score representing the suitability for each file is calculated and stored in a memory (not shown). To do.

  The experience score calculation unit 104 experienced from the operation history information in the operation history database 109 based on the identifier that is obtained through the search management unit 102 and specifies the method for calculating the experience score for the search request. An experience score representing the strength for each file operated on the PC is calculated and stored in a memory (not shown).

  The integrated score calculation unit 105 uses the operation history information on the basis of an identifier that specifies a method for calculating the integrated score for the search request acquired via the search management unit 102 (see FIG. An integrated score is calculated by integrating the file fitness score for each file stored in the memory (not shown) and the experience score for each file stored in the memory (not shown) of the experience score calculation unit 103, and the memory (see FIG. (Not shown).

  The all-page search unit 106 searches the external search engine 300 for an arbitrary range of information including files that are not viewed by the user as a population, and transfers the search result to the similarity calculation unit 107.

  The similarity calculation unit 107 acquires a file as a result of searching as a population the arbitrary range information including a file that is not viewed by the user from the all-page search unit 106, and stores the file as a search result file (URL). Based on this, the Web server 400 is accessed, the result is acquired, the similarity between the result and a file browsed in the past is calculated, and stored in a memory (not shown).

  The display priority calculation unit 108 uses the display priority adjustment information that specifies the method of calculating the display priority from the similarity and the integrated score, which is acquired via the search management unit 102, and uses the display priority adjustment information of the similarity calculation unit 107. The display priority is calculated from the similarity stored in the memory (not shown) and the integrated score stored in the memory (not shown) of the integrated score calculation unit 105.

  Next, the operation in the above configuration will be described.

  First, a method in which the operation history acquisition unit 101 acquires an operation history will be described.

  As illustrated in FIGS. 4 and 5, the operation history DB 109 includes a Web operation history table, an IME input table, a print table, a Web file table, and a mail file table.

  The Web operation history table (FIG. 4A) includes items of File-Key (information body hash value (MD5)), browsing start time, and browsing end time.

  The IME input table (FIG. 4B) is composed of items of occurrence time and input contents.

  The print table (FIG. 4C) is composed of items of occurrence time, file type, and file-key.

  The web file table (FIG. 5 (d)) is published on the Internet as Content-Key (hash value of information body (MD5)), file name storing the web page seen in the past, author, title, update date. It consists of the items of the URL that has been.

  FIG. 6 is a diagram showing an operation history information registration operation in the operation history acquisition unit according to the embodiment of the present invention.

・ Web:
The operation history acquisition unit 101 acquires (hooks) a window message that is globally communicated on the Windows (registered trademark) OS when a keyboard event, a mouse event, or an active window state change event occurs. (Step 101), it is specified which program is being operated by the user (Step 102). You can also get information about which windows are active or inactive. In the present embodiment, description will be made assuming that the Internet Explorer (registered trademark) (hereinafter referred to as IE) of Micro Soft Corporation is used.

  Next, it communicates with the IE through inter-program communication (step 103), and acquires the URL of the home page currently displayed on the IE window and the displayed information body (in many cases, an HTML file) (step 104). ).

  Next, the MD5 hash value of the information body is calculated, and this value is set as a Content-key for uniquely identifying the content (step 105).

  Next, the information body is stored in the past operation file storage unit 111 of the PC as “file seen in the past” (step 106).

  Next, the activation time of the window included in the window message is estimated to be the browsing start time of the URL (step 107), and the browsing start time and the MD5 hash value of the information body are stored in the operation history DB 109 as the Web operation history. It is registered in the table (FIG. 4A) (step 108).

  Next, the operation history acquisition unit 101 uses the analysis result of the information body and the acquired information, and the author, title, update date, and MD5 hash of the information body, which are attribute information of the information body. The value, the URL published on the Internet, and the file name (full path) storing the information body are stored in the Web file table (FIG. 5D) (step 109). After that, when the operation history acquisition unit 101 receives a window deactivation message included in the window message (step 110), the operation history acquisition unit 101 estimates that time is the browsing end time of the URL (step 111). The browsing end time is registered in the Web operation history table (FIG. 4A) of the history DB 109 (step 112).

・ E-mail:
By acquiring (hooking) the window message, the IME input is stored in the IME input table (FIG. 4B) in the operation history DB together with the generation time.

·printing:
In response to an inquiry to the printer spool, the print event is stored in the print table (FIG. 4C) of the operation history DB 109 as an operation history together with the generation time.

  Next, a procedure in which the search device 100 performs a search will be described. Hereinafter, it is assumed that the operation history DB 109 accumulated as described above is used.

  FIG. 7 is a diagram showing a search operation in one embodiment of the present invention.

  The search management unit 102 receives the search keyword, display priority calculation method designation information (β = 0.6), and integrated score calculation method identifier (a: “prior author = Taro: other = default”, b) from the information search terminal 200. : “Browsing time of 5 minutes or longer; otherwise = default”, c: “(experience weight coefficient α, file matching weight coefficient 1−α) = (0.7, 0.3)”) is received (step 201) ).

  The search management unit 102 searches the operation history DB 109 to obtain a list of content-keys and file paths in which files seen in the past and mails seen in the past are stored (step 202).

  The search management unit 102 sends the search keyword, content-key, file path list, and integrated score calculation method identifier a “prior author = taro: other = default” to the file search unit 103 (step 203).

  The file search unit 103 uses the search keyword and the list of file paths to search the past operation file storage unit 111 for files that have been viewed in the past and mails that have been viewed in the past (step 204), and calculates a file suitability score (step 204). Step 205). At this time, it is checked whether or not the file is “Taro” by the author or whether it is a mail by accessing 110 with the file attribute, and the file conformity score of the file whose author is “Taro” is doubled. In addition, the file matching score of a file that does not include a search keyword (does not hit the search) and mail is “0”. The file search unit 103 transmits a list of pairs of Content-key and file matching score to the search management unit 102 (step 206).

The search management unit 102 transmits to the experience score calculation unit 104 the Content-key and the integrated score calculation method identifier b “browse the viewing time of 5 minutes or longer: otherwise = default” (step 207). The experience score calculation unit 104 accesses the operation history DB 109, calculates the browsing time length of a file browsed in the past (step 208), and calculates the experience score (step 209). This calculation is performed using the formula shown in FIG. However, the viewing time of 5 minutes or more is rounded down to 5 minutes. And a list of pairs of Content-key, experience score and browsing period (for example, “(Content-key browsing monitoring, browsing accommodation time, experience score) = (Fasfasfafadfasdfga, 2004/09/01/09: 11: 55,2003 / 09/01/09: 12: 45,170)
Is returned to the search management unit 102 (step 210).

  The search management unit 102 is a list of pairs of Content-key, experience score, and file match score, calculation method identifier = “(experience weight coefficient α, file match weight coefficient 1−α) = (0.7, 0.3) ] Is transmitted to the integrated score calculation unit 105 (step 211).

  The integrated score calculation unit 105 is a list of pairs of Content-key, experience score, and file match score, calculation method identifier = “(experience weight coefficient α, file match weight coefficient 1−α) = (0.7, 0.3 )] Is used to calculate the integrated score (step 212). For the calculation of the integrated score, a sum of a value obtained by multiplying the experience score by α = 0.7 and a value obtained by multiplying the file matching score by (1−α) = 0.3 is calculated as an integrated score. The integrated score time function is an index indicating how deeply a file matching a keyword was experienced at an arbitrary time. The integrated score calculation unit 105 transmits a list of content-key and integrated score pairs to the search management unit 102 (step 213).

  The search management unit 102 transmits the search keyword to the all page search unit 106 (step 214).

  The all-page search unit 106 is an external search engine 300 (for example, goo (registered trademark), Google (registered trademark)) that can search all the searchable files on the Internet including files that the user has not browsed in the past. , Etc., and the top 100 URLs, summary sentences, and titles that match the search keyword are acquired (step 215).

  The all-page search unit 106 transmits the URL, summary sentence, and title that match the search keyword to the search management unit 102 (step 216).

  The search management unit 102 transmits the URL (for 100 cases), the content-key of the file browsed in the past, and the file name (full path) in which the information body of the content-key is stored to the similarity calculation unit 107 ( Step 217).

  The similarity calculation unit 107 accesses the Web server 400 using the URL (for 100 cases) and acquires the file (for 100 cases) indicated by the URL (step 218). Next, the text similarity between each file and each content-key file browsed in the past is calculated (step 219). This can be realized by existing techniques such as using the cosine similarity of word vectors appearing in a document.

  Next, the URL similarity (= URL similarity) between the file and each Content-key file browsed in the past is calculated (step 220). For example, the URL similarity is 0.5 points if the URLs completely match and 0.5 points if only the site is the same, and 0.5 points if the URL site and the part of the path below it are the same. Points, 0 points for other sites.

  The similarity is the sum of the text similarity and the URL similarity.

The similarity calculation unit 107 includes 100 URLs and each Content-key viewed in the past.
The similarity to the file is transmitted to the search management unit 102 (step 221).

  The search management unit 102 displays the similarity between 100 URLs and Content-key files browsed in the past, a list of Content-keys, and display priority calculation method designation information (β = 0.6). It transmits to the priority calculation part 108 (step 222).

  The display priority calculation unit 108 calculates the display priority from the search management unit 102 and the list of similarities and content-keys between 100 URLs from the similarity calculation unit 107 and the content-key files browsed in the past. The method designation information (β = 0.6) is received. The display priority calculation unit 108 calculates the display priority of the content-key file browsed in the past and displayed on each URL (step 223). The calculation method is shown in FIG. The display priority calculation unit 108 transmits the URL and the priority of the content-key file browsed in the past for each URL to the search management unit 102 (step 224).

  For each URL, the search management unit 102 selects three files with high display priority from among the content-key files browsed in the past, and displays them in order of display priority for each URL. A search result display file is created (step 225). Next, the search result display file is transmitted to the information search terminal 200 (step 226). A screen image displayed on the information search terminal 200 is shown in FIG.

  A user who operates the information search terminal 200 can input a search keyword when making a search request. Also, the content similarity and the weight of the experience score, which is the degree experienced in the past, are adjusted, and each URL has a high display priority for actual display in the content-key file browsed in the past. It is possible to operate a bar for adjusting display priority information for changing three files and a bar for adjusting the weight of the experience score for calculating the integrated score and the file matching score.

  The user's operation history may be acquired from the operation history acquisition unit 101 or may be registered by a registration request from an application of another user terminal such as Internet Explorer (registered trademark). In some cases, the user may register from a Web server that requested information to be viewed.

  Note that the operation of the search device 100 in the above embodiment can be constructed as a program, installed on a single server, an existing server, a user terminal, or distributed via a network.

  In addition, the constructed program is stored in a portable storage medium such as a hard disk device or a flexible disk / CD-ROM, and installed in a single server, an existing server, or a user terminal, or It is possible to distribute.

  The present invention is not limited to the above-described embodiment, and various modifications and applications can be made within the scope of the claims.

  The present invention can be applied to a technique for searching for operation contents of a user's PC.

It is a figure for demonstrating the principle of this invention. It is a principle block diagram of this invention. It is a block diagram of the search device in one embodiment of this invention. It is an example (the 1) of the table of operation history DB in one embodiment of this invention. It is an example (the 2) of the table of operation history DB in one embodiment of this invention. It is a figure which shows the registration operation | movement of the operation history information in one embodiment of this invention. It is FIG. (1) which shows the search operation in one embodiment of this invention. It is FIG. (2) which shows the search operation in one embodiment of this invention. It is an example of the experience score calculation in one embodiment of this invention. It is a figure which shows the display priority calculation method in one embodiment of this invention. It is a screen image in one embodiment of the present invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 Information search device 101 Operation history acquisition part 102 Search management means, Search management part 103 File search means, File search part 104 Experience score calculation means, Experience score calculation part 105 Integrated score calculation means, Integrated score calculation part 106 All page search means , All page search unit 107 similarity calculation unit, similarity calculation unit 108 display priority calculation unit, display priority calculation unit 109 operation history DB (database)
110 File attribute DB (database)
111 Past operation file storage unit 112 Score calculation means 200 Information retrieval terminal (IE)
300 External search engine 400 Web server

Claims (6)

An information search method for searching operation history information,
When receiving a search request from the request source, the file search means searches for an information body (file) browsed in the past by a user who matches the search request based on the search request;
Based on the storage location information of the file, the score calculation means indicates the strength that the user has experienced using the file suitability score indicating the degree of suitability for the search request for each file or the operation history information in the operation history database. A score calculation step of calculating an experience score represented for each file operated on the PC, and calculating an integrated score obtained by integrating the file suitability score and the experience score;
Similarity calculation means, a similarity calculation step for obtaining a similarity between a file as a result of searching information in an arbitrary range including a file that the user has not browsed as a population, and a file browsed in the past,
An output step for the search management means to output a search result based on the display priority obtained from the similarity and the integrated score;
An information retrieval method characterized by: Receiving from the requester an identifier specifying a method for calculating a file conformity score to the search request, the experience score, and the integrated score;
In the score calculating step,
According to the identifier, changing the calculation method of the file suitability score, the experience score, and the integrated score for the search request, and calculating the file suitability score, the experience score, and the integrated score;
The information search method according to claim 1, wherein: Receiving display priority adjustment information for designating a method for calculating the display priority from the similarity and the integrated score from the requester;
Calculating display priority using the display priority adjustment information;
The information search method according to claim 1, wherein: The operation history acquisition means
Obtaining a history of operating the user's PC using communication with a message message on the PC and a program operating on the PC;
Obtaining an identifier for identifying information viewed by the user from a history of operating the user's PC;
Based on an identifier for identifying information browsed by the user, obtaining an information body browsed in the past by the user, and storing it in the operation history database as a file;
Obtaining attribute information of the information body and storing it in a file attribute information database as file attribute information;
The information search method according to claim 1, wherein: An information retrieval device for retrieving operation history information,
Operation history information, which is a history of operating a user's personal computer (PC), an operation history database,
Based on a search request acquired from a request source, a file search means for searching an information body (file) browsed in the past by a user who matches the search request based on the search request;
Based on the storage location information of the file, the file conformity score indicating the degree of conformity to the search request for each file or the strength experienced by the user using the operation history information of the operation history database is operated on the PC. A score calculation means for calculating an experience score represented for each file and calculating an integrated score obtained by integrating the file conformity score and the experience score;
Similarity calculation means for obtaining a similarity between a file as a result of searching an arbitrary range of information including a file that is not browsed by the user and a file browsed in the past,
Based on the display priority obtained from the similarity and the integrated score, an output for outputting a search result;
An information retrieval apparatus comprising: On the computer,
5. A computer-readable recording medium storing an information search program for executing each step of the information search method according to claim 1.

Images