JP2004206211A - Information retrieval device, information retrieval method, program and storage medium storing the program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To reduce a user's burden by dispensing with a keyword input by the user and to obtain a reliable retrieval result by the use of a proper keyword. <P>SOLUTION: The retrieval server 5 (information retrieval device) comprises a retention means for patrolling a network on a regular base or irregular base according to a prescribed site list, downloading a homepage present in the network for every prescribed category, and retaining a character string contained in the homepage in association with the view frequency of the homepage and the address of the homepage; a retrieval means for retrieving, according to a predetermined keyword dictionary, information to be provided to a user from the information retained in the retaining means for every category; and a disclosure means for generating a content listing the information retrieved by the retrieval means or its link information and disclosing the content to the network. <P>COPYRIGHT: (C)2004,JPO&NCIPI

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to an information search device, an information search method, a program, and a recording medium storing the program. More specifically, an information search device, an information search method, a program, and the like that search for information considered important to a user from various types of information existing on a network such as the Internet or an intranet, and present the information to the user. The present invention relates to a recording medium storing a program.
[0002]
[Prior art]
Huge amounts of information are flooding networks such as the Internet and intranets, especially the Internet, which is a global wide-area network, and it is extremely difficult to find information that is important to users from the information. It is a technique of. For this reason, service sites called “search sites” are provided on the Internet, and users use the search sites to find desired information. These search sites take a so-called “pull-type” service form. That is, the pull-type search site receives a keyword from the user, searches for information that matches the keyword, and provides the user with the information.
[0003]
The pull-type search service is useful when a user wants to know specific information, that is, when he / she uses it spontaneously. However, when it is desired to know information that changes every moment, for example, information such as news, the pull-type search service is inconvenient. This is because each time a keyword must be entered. In addition, if the keyword is inappropriate, accurate information cannot be obtained.
[0004]
Therefore, by performing user registration in advance, the search service provider generates a dedicated screen listing information dedicated to the user (for example, a dedicated news screen covering the area of interest of the user). A so-called “push-type” search service that distributes to a user's client terminal is used.
[0005]
As a conventional push-type information search method, a keyword is extracted from registration information at the time of user registration, and based on the keyword, a variety of information existing on a network such as the Internet or an intranet is used. There is known an apparatus that searches for information that is considered to be available and distributes the information to a user in a push type (for example, see Patent Document 1).
[0006]
[Patent Document 1]
JP-A-2000-298677 ([0302] to [0327], FIG. 105)
[0007]
[Problems to be solved by the invention]
However, the conventional push-type information search method requires a user to input a keyword at the time of user registration, and in this point (keyword input), it has the same disadvantage as the pull-type search service. I have. That is, there is a problem in that the input of the keyword is troublesome and the input keyword may be inappropriate, so that the expected search accuracy cannot always be obtained.
[0008]
Therefore, it is an object of the present invention to reduce the burden on the user by eliminating the need for a user to input a keyword, and to obtain a highly reliable search result using an appropriate keyword.
[0009]
[Means for Solving the Problems]
The information search device according to the present invention is a patrol information holding unit that holds information of a network to be patrolled, and a patrolling unit that patrols the network periodically or irregularly according to the content held in the patrolling information holding unit. Download means for downloading homepages existing in the network for each predetermined category by patrol by the patrol means, character strings included in the homepage acquired by the download means, view frequency of the homepage, and address of the homepage , A keyword dictionary holding means for holding a keyword dictionary for each category, and a user provided from the information held in the holding means according to the keyword dictionary held in the keyword dictionary holding means. Information to be classified by category A search to find means to generate content that list the retrieved information or the link information by the retrieving means, characterized in that a public means for publishing the content to the network.
[0010]
According to the present invention, a homepage existing on a network is downloaded for each predetermined category, and a character string included in the homepage, a view frequency of the homepage, and an address of the homepage are stored in association with each other. From the information, information to be provided to the user is searched according to the keyword dictionary, and the content including the searched information or its link information is generated and made public on the network.
[0011]
Therefore, by preparing an appropriate keyword dictionary, information that is attracting attention at that time can be collected for each category and can be provided to the user in a push type. As a result, keyword input by the user becomes unnecessary. The user's burden can be reduced, and a highly reliable search result can be obtained using an appropriate keyword.
[0012]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the following description, the specification or examples of various details and examples of numerical values, character strings, and other symbols are merely reference for clarifying the idea of the present invention, and all or some of them Obviously, the idea of the invention is not limited. In addition, well-known techniques, well-known procedures, well-known architectures, and well-known circuit configurations (hereinafter, “known matters”) will not be described in detail, but this is also for the purpose of simplifying the description, It does not intentionally exclude all or some of the matters. Such a well-known matter can be known to those skilled in the art at the time of filing the present invention, and is naturally included in the following description.
[0013]
FIG. 1 is an overall system configuration diagram of an embodiment. In FIG. 1, a plurality of WWW (World Wide Web) servers 2a to 2d are connected to a network (hereinafter, referred to as "Internet") 1 such as the Internet or an intranet. These WWW servers 2a to 2d are given addresses called URLs (Uniform Resource Locators), and can be accessed from a plurality of user terminals 3a to 3d also connected to the Internet 1 by specifying each address. It has become.
[0014]
That is, the WWW servers 2a to 2d communicate with a document (in many cases, such as "index.html" or ".html") described in a structured language such as HTML (Hyper Text Markup Language) or XML (extensible Markup Language). And a variety of contents linked to the top page. Hereinafter, these will be collectively referred to as “homepage” for convenience. The user inputs a desired homepage address using homepage browsing software (so-called browser software) of his own terminal (any of the user terminals 3a to 3d), downloads the homepage to his own terminal, and browses it.
[0015]
Here, only four WWW servers 2a to 2d on the drawing are shown. This is in order to avoid drawing congestion, and in fact, especially in the case of the Internet, a huge number of units on a global scale, and therefore, the amount of home pages becomes extremely large. For this reason, the above-mentioned browsing procedure, that is, the procedure of browsing by entering the address of the home page using browser software is effective for browsing a home page whose address is known, but browsing a home page whose address is unknown. Not applicable to
[0016]
The search service site 4 connected to the Internet 1 is provided for convenience in browsing a homepage whose address is unknown, and particularly provides a push-type information search service. The search service site 4 includes a search server 5 (information search device) connected to the Internet 1.
[0017]
FIG. 2 is a diagram illustrating a configuration of the search server 5. The search server 5 includes a CPU 10 (a traveling unit, a downloading unit, a searching unit, a publishing unit), a RAM 11, a disk controller 12, a disk unit 13 (a traveling information holding unit, a holding unit, a keyword dictionary holding unit), a display controller 14, and a display unit. 15, a keyboard controller 16, a keyboard device 17, a pointing device device 18, a communication control device 19, a main bus 20, a bus interface 21, an internal bus 22, and the like.
[0018]
The search server 5 implements a predetermined search service having the following functions by loading basic software and application software stored in the disk device 13 in advance into the RAM 11 and executing the same by the CPU 10. It is provided on the Internet 1 through the Internet.
[0019]
(1) Member (user) management function
The search server 5 holds and manages specific information such as the name, address, and contact information of the user who is to receive the search service in a predetermined database file of the disk device 13. When the service is provided for a fee, payment information for the charging (for example, in the case of a card payment, information such as a card company name, a card number, and a card expiration date) is further held and managed. If so, it also retains and manages other information useful for marketing research and sales activities, for example. Further, a site list and a category-specific keyword dictionary, which will be described later, are held and managed for each user.
[0020]
(2) Internet information collection function
The search server 5 automatically or manually patrols the Internet 1 periodically or irregularly according to a predetermined site list (see FIG. 3) held for each user, and downloads home pages existing in the Internet 1 for each predetermined category. Then, a character string included in the home page, a view frequency of the home page, and an address of the home page are associated with each other and stored for each user.
[0021]
(3) Information search function
The search server 5 searches, for each category, information to be provided to the user from the information stored in (2) according to a category-specific keyword dictionary described later (see FIG. 4).
[0022]
(4) Content generation and distribution function
The search server 5 generates contents (contents in a document format described by a structured language such as HTML) listing information searched for in (3) or link information thereof, and publishes the contents to the Internet 1. . When only the registered user is permitted to view the content (that is, when anonymous disclosure is not permitted), the content is protected by a password or the like.
[0023]
(4) Billing function
When browsing the content for a fee, the search server 5 performs a charging process for the registered user according to a pay-as-you-go system such as the number of times of browsing or a fixed amount such as a monthly usage fee collection.
[0024]
FIG. 3 is a conceptual diagram of the structure of the site list. The site list 20 includes a category number field 21, a category name field 22, a site number field 23, a site URL field 24, a view frequency value field 25, an applied keyword dictionary field 26, and the like.
[0025]
... Are stored in the category No. field 21, and an appropriate category (for example, “news”, “news”, “Sports”, “entertainment”, etc.) are stored. The site number field 23 stores serial numbers "1, 2, 3, ..." for each category, and the site URL field 24 stores the URL of each site. The view frequency value field 25 stores a value (for example, the number of times of browsing) representing the browsing frequency (view frequency) of the top page of each site, and the applied keyword dictionary field 26 is applied to each category. (E.g., “News keyword dictionary”, “Sports keyword dictionary”, “Entertainment keyword dictionary”, etc.) are stored (to store the entity of the dictionary or to store link information to the dictionary). Means to store.)
[0026]
FIG. 4 is a conceptual diagram of a category-based keyword dictionary. Here, three keyword dictionaries of a news keyword dictionary 27, a sports keyword dictionary 28, and a performing arts keyword dictionary 29 corresponding to the site list 20 are shown. Each of these keyword dictionaries 27 to 29 has a keyword field 27a and an evaluation value field 27b. The keyword field 27a stores various words, personal names, place names, general nouns, and the like (KW1, KW2, KW3, KW4, KW5,... In the figure for convenience), and an evaluation value field 27b. Stores a value indicating the importance of the keyword in the category (for example, a large value is assigned to a keyword related to a highly newsworthy case or a person).
[0027]
FIGS. 5 and 7 are conceptual flowcharts showing the main parts of the application software executed by the CPU 10 of the search server 5. This flowchart is repeatedly executed at predetermined time intervals. This flowchart includes two nested loops, as will be apparent from the following description. The first loop is a "category loop", and the second loop is a "site loop" repeated within the first loop. The variable i is a loop counter of a category loop, and the variable j is a loop counter of a site loop.
[0028]
In this flowchart, first, an initial value “1” is set to a variable i and a variable j, and the maximum value of the category number of the site list 20 is set to a variable imax that defines the maximum number of times of the category loop ( Step S10). Here, the maximum value of the category number is a value corresponding to the total number of categories registered in the site list 20. For example, as described above, when the serial number of “1, 2, 3,...” Is stored in the category No., it is the last value of the serial number. Hereinafter, for convenience of explanation, the maximum value of the category number of the site list 20 is set to “3”. Therefore, in this case, “3” is set to the variable imax that defines the maximum number of times of the category loop.
[0029]
Next, the maximum value of the site No. of the category No. (i) is set in a variable jmax that defines the maximum number of site loops (step S11). Here, the category No. (i) means “category No.” + (The value of the variable i at that time). That is, when the variable i = 1, the category No. (i) indicates the category No. = 1. In FIG. 3, the category No. = 1 is the category of "news", and the maximum value of the site No. of the category is, for example, "4" of the site URL "www.news4.com". Therefore, in this case, “4” is set to the variable jmax that defines the maximum number of site loops.
[0030]
Next, the user accesses the URL of the site No. (j) of the category No. (i) and downloads its top page (step S12). Here, the category No. (i) means “category No.” + (The value of the variable i at that time) as described above, and similarly, the site No. (j) is “the site No. "+ (The value of the variable j at that time). That is, when the variable i = 1 and the variable j = 1, the site No. (j) of the category No. (i) indicates the site No. = 1 of the category No. = 1. In FIG. 3, the site No. = 1 of the category No. = 1 is the first site in the category of "news", and the URL of the site is "www.news1.com". Therefore, in this case, the top page of “www.news1.com” is downloaded.
[0031]
Next, it is checked whether or not the downloaded top page includes a word (or word, etc.) that matches each keyword registered in the applicable keyword dictionary of the category No. (i) (step S13, step S14). . Here, the applicable keyword dictionary of the category No. (i) means a keyword dictionary for each category corresponding to the value of the variable i. For example, if the variable i = 1, the category No = 1 is the category of "news", and thus means the news keyword dictionary 27 (see FIG. 4). As described above, various keywords (KW1, KW2, KW3, KW4, KW5,...) Are registered in the news keyword dictionary 27 together with the evaluation value in the category (the keyword field 27a and the evaluation in FIG. 4). (See the value field 27b), and in these steps S13 and S14, whether or not those keywords (KW1, KW2, KW3, KW4, KW5,...) Are included in the downloaded top page. Check whether or not.
[0032]
If it is included (“YES” in step S14), the URL of the top page is registered in the keyword appearance table of category No. (i), and a predetermined keyword appearance symbol (black circle described later) is entered. Symbol).
[0033]
FIG. 6 is a conceptual diagram of the structure of the keyword appearance table 30. The keyword appearance table 30 is provided for each category. For example, when "3" is set in a variable imax that defines the maximum number of times of the category loop, the total number of categories is "3". One keyword appearance table 30 is provided. For convenience of explanation, the illustrated keyword appearance table 30 is for the category of “news”. This keyword appearance table 30 is used to store each keyword (KW1, KW2, KW3, KW4, KW5,...) Registered in the news keyword dictionary 27 (see the keyword field in FIG. 4). A keyword field 31, an evaluation value field 32 for storing an evaluation value for each keyword registered in the news keyword dictionary 27 (see the evaluation value field 27b in FIG. 4), a URL registration field 33, a keyword And an appearance evaluation value field 34.
[0034]
The URL registration field 33 has a number of URL registration units 33a to 33h corresponding to the value of the variable jmax that defines the maximum number of site loops in the category. The URL registration units 33a to 33h are sequentially numbered from the left corresponding to the value of the variable (j). For example, when the variable i = 1 and the variable j = 1, the above-described step is performed. If the determination result in S14 is “YES”, that is, the keywords (KW1, KW2, KW3, KW4, KW5) are included in the top page downloaded from the first site “www.news1.com” in the “news” category. ,...), The URL of the top page (“www.news1.com”) is registered in the URL registration unit 33a on the left end.
[0035]
In the illustrated example, “URL1”, “URL2”, “URL3”,... Are stored in the respective URL registration units 33a to 33h, but these are stored in actual URLs (for example, “www”). .News1.com ”). The numerical values (“78”, “83”, “92”, “64”,...) Appended to the lower part of each of the URL registration sections 33a to 33h are “URL1,” “URL2,” “ URL3 ",... Are the values looked up from the view frequency value field 25 of the site list 20 in FIG.
[0036]
Here, black circles (•) in the figure are keyword appearance symbols (may be read as keyword appearance flags). For example, focusing on the row of KW1, five black circles are set in the row, and the positions of the black circles correspond to the columns of URL1 to URL4 and URL8, respectively. Therefore, in this case, it is specified that the keyword (KW1) is included in each of the top pages of URL1 to URL4 and URL8. In the keyword appearance evaluation value field 34, the answer of the following formula (1) (hereinafter referred to as "keyword appearance evaluation value") is set.
[0037]
Keyword appearance evaluation value = (Evaluation value of the keyword) × (Number of black circles in the keyword line)... ▲ 1 ▼
For example, since the evaluation value of KW1 is “511” and the number of black circles in the KW1 row is “5”, the answer of the equation (1) is “511 × 5 = 2,555”. Therefore, in this case, the appearance evaluation value of the keyword (KW1) in the category of “news” is “2,555” as illustrated.
[0038]
Similarly, since the evaluation value of KW2 is “367” and the number of black circles in the KW2 row is “5”, the answer of Expression (1) is “367 × 5 = 1,835”. Therefore, in this case, the appearance evaluation value of the keyword (KW2) in the category of “news” is “1,835” as illustrated.
Similarly, since the evaluation value of KW3 is “403” and the number of black circles in the KW3 row is “4”, the answer of Expression (1) is “403 × 4 = 1,612”. Therefore, in this case, the appearance evaluation value of the keyword (KW3) in the category of “news” is “1,612” as illustrated.
Similarly, since the evaluation value of KW4 is "301" and the number of black circles in the KW4 row is "4", the answer to the expression (1) is "301 × 4 = 1,204". Therefore, in this case, the appearance evaluation value of the keyword (KW4) in the category of “news” is “1,204” as illustrated.
Similarly, since the evaluation value of KW5 is “192” and the number of black circles in the KW5 row is “4”, the answer of Expression (1) is “192 × 4 = 768”. Therefore, in this case, the appearance evaluation value of the keyword (KW5) in the category of “news” is “768” as illustrated.
Similarly, since the evaluation value of KW6 is “219” and the number of black circles in the row of KW6 is “3”, the answer of Expression (1) is “219 × 3 = 657”. Therefore, in this case, the appearance evaluation value of the keyword (KW6) in the category of “news” is “657” as illustrated.
Similarly, since the evaluation value of KW7 is “97” and the number of black circles in the KW2 row is “6”, the answer of Expression (1) is “97 × 6 = 582”. Therefore, in this case, the appearance evaluation value of the keyword (KW7) in the category of “news” is “582” as illustrated.
Similarly, since the evaluation value of KW8 is “530” and the number of black circles in the KW8 line is “1”, the answer of the equation (1) is “530 × 1 = 530”. Therefore, in this case, the appearance evaluation value of the keyword (KW8) in the category of “news” is “530” as illustrated.
[0039]
Thus, the keyword appearance table 30 of one category (the category of “news” when the variable i = 1) is generated. When the keyword appearance table 30 of the category is completed, the variable j is sequentially set. (Step S16), and the value of the variable j exceeds the variable jmax. That is, when the site loop in the category is completed (“YES” in step S17).
[0040]
When the site loop in the category is completed (“YES” in step S17), the variable i is counted up (step S18), and it is determined whether the variable i after the count exceeds the variable imax (step S18). S19). If not exceeded, it is determined that there is a remaining category, and the processes after step S11 are repeatedly executed. As a result, the keyword appearance tables 30 for other categories (such as "sports" and "entertainment") are generated in the same manner as described above. The number of times of the category loop is imax. The number of site loops is the number of sites registered in each category for each category.
[0041]
When the keyword appearance tables 30 for all categories are completed, push content is generated based on the keyword appearance tables 30 and published to the Internet 1 (step S20), and the flowchart ends.
[0042]
FIG. 8 is a diagram illustrating an example of the push content 35 (content). This push content 35 represents a state displayed on a user terminal (user terminals 3a to 3d). The push content 35 is a document described in a structured language such as HTML or XML, and, for example, currently attracts attention for each category (in the illustrated example, “news”, “sports”, “entertainment”, etc.). The most important information is provided with links 36-38. In the illustrated example, matching keywords (for example, KW1, KW2, KW3,...) Are clearly displayed on the screen as display modes of the links 36 to 38, but this is a simple example. It is just that. The URL of the link destination may be displayed together with the keyword, or a part of a character string or an image included in the top page of the URL may be displayed. Alternatively, the above character strings and the like may be displayed in a pop-up format when the links 36 to 38 are focused.
[0043]
As described above, the information (links 36 to 38) to be displayed on the push content 35 is selected based on the keyword appearance table 30, as described above.
[0044]
<Selection by a single keyword>
In the keyword appearance table 30 of FIG. 6, the maximum value of the keyword appearance evaluation value is “2,555” corresponding to KW1 in the first row, and the appearance number (the number of black circle symbols) of KW1 is the highest in the category. Therefore, when a selection is made using a single keyword, a link may be provided to a URL (URL1 to URL4 in the example shown) that publishes a top page including this keyword (KW1).
[0045]
However, in the case of the illustrated example, the number of URLs is small (four URLs, URL1 to URL4), so that there is no problem in displaying on the push content 35. However, in some cases, a large number of URLs may be hit. In this case, it is difficult to find useful information on the user's terminal. For example, only the hit URL having the highest view evaluation value (URL3) is displayed on the push content 35, or the view evaluation is performed. Some of the higher values may be displayed in the push content 35.
[0046]
<Selection by multiple keywords>
Generally, when finding desired information from a large amount of information, an AND search for a plurality of keywords is performed. For example, the search condition is specified as “including all of KW1, KW2, and KW3”. Such a condition designation is called “and search” or “and combination” in the field of a database, and is expressed in a format of “search keyword” and “search keyword”. In the present invention, when such an and search is applied, the following may be performed.
[0047]
In the keyword appearance table 30 of FIG. 6, for example, when performing an AND search of KW1 to KW4, URLs including all of KW1 to KW4 are URL1, URL3, and URL4. Therefore, by transforming the above formula (1),
Keyword appearance evaluation value = (sum of evaluation values of each keyword) x (number of URLs including each keyword) ... 2
Then, the sum of the evaluation values of KW1 to KW4 is “511 + 367 + 403 + 301 = 15882”, the number of URLs including each keyword (KW1 to KW4) is “3” of URL1, URL3, and URL4, and the answer of equation (2) Is “1582 × 3 = 4,746” and the maximum keyword appearance evaluation value is obtained. Therefore, a URL that publishes a top page including these keywords (KW1 to KW4) (URL1, URL3 in the example shown) And URL 4).
[0048]
However, in the case of the example shown in the figure, since there are only a small number of URLs (three URL1, URL3 and URL4), there is no problem in displaying on the push content 35. However, in some cases, a large number of URLs may be hit. In this case, it is difficult to find useful information on the user's terminal. For example, only the hit URL (URL1, URL3, and URL4) having the highest view evaluation value (URL3) is pushed content 35. Or some of the higher-ranked view evaluation values may be displayed on the push content 35.
[0049]
As described above, according to the present embodiment, the search server 5 searches for information that is currently attracting attention for each category such as “news”, “sports”, and “entertainment”, and searches the information. Push content 35 with links 36 to 38 can be generated and made public on the Internet 1. Therefore, the user can use his / her terminal (user terminals 3a to 3d) to browse the push content 35 whenever he wants, and can touch the latest useful information without inputting a keyword each time. The effect of can be obtained.
[0050]
【The invention's effect】
According to the present invention, a homepage existing on a network is downloaded for each predetermined category, and a character string included in the homepage, a view frequency of the homepage, and an address of the homepage are held in association with each other. From the held information, information to be provided to the user is searched according to a predetermined keyword dictionary, and a content including the searched information or the link information is generated and made public on the network.
[0051]
Therefore, by preparing an appropriate keyword dictionary, information that is attracting attention at that time can be collected for each category and can be provided to the user in a push type. As a result, keyword input by the user becomes unnecessary. The user's burden can be reduced, and a highly reliable search result can be obtained using an appropriate keyword.
Further, according to a preferred aspect of the present invention, a management unit for managing at least one of the tour information holding unit, the holding unit, and the keyword dictionary holding unit is provided for each user, and the content generation and disclosure service is provided by the user. This can be performed for each user, and a detailed service corresponding to each user can be realized.
[Brief description of the drawings]
FIG. 1 is an overall system configuration diagram of an embodiment.
FIG. 2 is a diagram showing a configuration of a search server 5.
FIG. 3 is a conceptual diagram of the structure of a site list.
FIG. 4 is a conceptual diagram of a keyword dictionary.
FIG. 5 is a conceptual flowchart (part 1) illustrating a main part of application software executed by the CPU 10 of the search server 5.
FIG. 6 is a conceptual diagram of the structure of a keyword appearance table 30.
FIG. 7 is a conceptual flowchart (part 2) illustrating a main part of application software executed by the CPU 10 of the search server 5.
FIG. 8 is a diagram showing an example of a push content 35.
[Explanation of symbols]
1 Internet (network)
5. Search server (information search device)
10 CPU (touring means, download means, search means, disclosure means)
13 Disk device (tour information holding means, holding means, keyword dictionary holding means)
20 Site List
27 Keyword Dictionary for News (Keyword Dictionary by Category)
28 Sport Keyword Dictionary (Keyword Dictionary by Category)
29 Keyword Dictionary for Entertainment (Keyword Dictionary by Category)
35 Push Content (Content)

Claims (5)

Traveling information holding means for holding information on a network to be visited,
In accordance with the contents held in the tour information holding means, a tour means that tours the network periodically or irregularly,
Download means for downloading home pages existing in the network for each predetermined category by patrol by the patrol means;
Holding means for associating a character string included in the homepage acquired by the download means with a view frequency of the homepage and an address of the homepage,
Keyword dictionary holding means for holding a keyword dictionary for each category;
Searching means for searching, for each category, information to be provided to the user from the information held in the holding means according to the keyword dictionary held in the keyword dictionary holding means;
An information search device, comprising: a publishing unit that generates a content listing information searched by the search unit or link information thereof, and publishes the content to a network. 2. The information retrieval apparatus according to claim 1, further comprising a management unit for managing at least one of the tour information holding unit, the holding unit, and the keyword dictionary holding unit for each user. According to a predetermined site list, periodically or irregularly patrols the network, downloads a homepage existing in the network for each predetermined category, a character string included in the homepage, a view frequency of the homepage, and A holding process for holding the homepage address in association with the address,
A search step of searching, for each category, information to be provided to the user from the information held in the holding step, according to a predetermined category-specific keyword dictionary;
A publishing step of generating a content listing information searched for in the search step or link information thereof, and publishing the content to a network. A program for causing a computer to execute a predetermined processing function,
The processing function includes:
According to a predetermined site list, periodically or irregularly patrols the network, downloads a homepage existing in the network for each predetermined category, a character string included in the homepage, a view frequency of the homepage, and A holding process for holding the homepage address in association with the address,
A search step of searching, for each category, information to be provided to the user from the information held in the holding step, according to a predetermined category-specific keyword dictionary;
A program for generating a content listing information searched for in the search step or link information thereof, and publishing the content to a network. A recording medium storing the program according to claim 4.

Images