JP2006236221A - Management server for web page retrieval - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a device retrieving Web pages related to a Web page even, in a case where no referencing is provided by a link or the like. <P>SOLUTION: An element information generation means 10 analyzes constituting elements of character information of every acquired Web page and generates Web page element information. The generated Web page element information is recorded for every Web page in a recording part 20. A receiving means 12 receives a retrieval request with a specified object Web page from a terminal device 4. An element information reading means 14 reads element information of the object Web page shown by the retrieval request from the recording part. A retrieval means 18 retrieves other Web pages having element information similar to the read element information of the object Web page from the recording part 20. A retrieval result transmitting means 16 transmits attribute information of retrieved Web pages such as URL to the terminal device 4 which is the requester. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a technique for finding a web page related to a target web page such as its own web page.

  The relationship between web pages on the Internet (bulletin board, blog, homepage, etc.) can be known by referring to how the links are pasted. For example, a search engine (such as google (trademark)) can be used to determine which other web page has a link to its own web page.

  Japanese Patent Application Laid-Open No. 2004-151561 discloses an apparatus that displays the reference relationship of articles in a bulletin board in a tree structure and can grasp the mutual relationship at a glance.

JP-A-9-106331

  However, the web page and the article must be associated with each other by technical means regardless of whether the search engine knows the link source or the method of displaying the relevance of the article as in Patent Document 1. I had to. Therefore, even if there is a similar web page as the content, it is difficult to find it if it is not related by a link or the like.

  An object of the present invention is to solve the above-described problems and to provide an apparatus and a method capable of searching for a web page related to a web page without being linked by a link or the like.

(1) (2) The management server device according to the present invention is a management server device for searching other web pages highly relevant to the target web page on the network, and accessing each web page. Character information acquisition means for acquiring character information constituting each web page, elements included in the acquired character information of each web page are extracted, web page element information is generated, and location specifying information of each web page is obtained. Element information generating means for recording in the recording unit in association with the included web page attribute information, receiving means for receiving a search request specifying the target web page sent from the terminal device, and target specified by the search request Element information reading means for reading out the web page element information of the web page from the recording unit, and referring to the recording unit, it is classified into the read target web page element information. It comprises a searching means for acquiring a Web page attribute information by searching the web page having a web page element information, and a search result transmitting means for transmitting a web page attribute information of the retrieved web page to the terminal device.

  Therefore, it is possible to search for other web pages having contents related to the target web page even if they are not linked by a link or the like.

(3) In the management server device according to the present invention, the element information generating means is based on the morpheme extracting means for extracting the morpheme as an element by analyzing the character information of each web page, and the appearance frequency of each extracted morpheme And a score calculation means for calculating a score for each morpheme and a morpheme table recording means for recording a score for each morpheme for each web page as a morpheme table in a recording unit. A search morpheme extraction means that extracts a predetermined number of morphemes from a plurality of morphemes that are page element information based on the score to obtain a search morpheme, and a morpheme table is given to each search page for each morpheme. There may be provided web page selection means for summing up the scores obtained and selecting similar web pages based on the sum.

  Therefore, the web page character information can be analyzed into morphemes and scored to accurately search related web pages.

(4) In the management server device according to the present invention, the score calculation means, based on the ratio of the number of applications in the web page of each morpheme extracted from the web page and the total number of morphemes in the web page, A score may be calculated.

  Therefore, the score for each morpheme can be calculated in consideration of not only the number of applications but also the weight of the entire application.

(5) In the management server device according to the present invention, the character information acquisition means and the element information generation means may perform their operations every predetermined period.

  Therefore, the processing burden on the management server device can be reduced.

(6) In the management server device according to the present invention, before the element information reading means and the search means receive the search request, the web page search is executed by using each web page as the target web page in advance. When the search result transmission means receives the search request from the terminal device, the web page attribute information of the web page similar to the target web page recorded in the recording unit is transmitted to the terminal device. Also good.

  Therefore, it is possible to quickly present related web page attribute information in response to a search request.

(7) In the management server device according to the present invention, the search result transmitting means may group each web page and transmit it to the terminal device based on the morpheme representing each searched web page.

  Therefore, what morpheme can provide what group of web pages is related.

(8) In the management server device according to the present invention, the web page searched by the search means is set as the second target web page, and the second page of the second target web page is read out from the recording unit. Element information reading means; and second search means for searching for a web page having web page element information similar to the read second target web page element information with reference to the recording unit, and the search result transmitting means The web page attribute information of the web page found by both the search means and the second search means may be transmitted to the terminal device.

  Therefore, the related web page can be further known about the related web page.

(9) In the management server device according to the present invention, the element information generation means generates and records the web page element information of each web page divided every predetermined period, and the search means uses the target web for each predetermined period. Based on the page element information, a related web page may be searched by specifying a period.

  Therefore, it is possible to search for a related web page by specifying each period to be converted to a time series of web pages.

(10) In the management server device according to the present invention, the web page is a blog.

  In general, by applying the present invention that can find related items to blogs that are frequently updated, usefulness is further increased.

(11) A web page search method according to the present invention is a web page search method for searching other web pages highly relevant to a target web page on a network by a computer, wherein the computer includes: Access each web page, get the text information that makes up each web page, extract the elements contained in the text information of each web page that was obtained, generate web page element information, and place specific information on each web page The web page element information of the target web page specified by the search request received from the terminal device is received in response to the search request specifying the target web page, recorded in the recording unit in association with the web page attribute information including Is read from the recording unit, is referred to the recording unit, and is similar to the read target web page element information. A web page having element information is searched to acquire web page attribute information, and the web page attribute information of the searched web page is transmitted to the terminal device.

  Therefore, it is possible to search for other web pages having contents related to the target web page even if they are not linked by a link or the like.

  In this invention, “web page” refers to content that can be browsed on the Internet. The concept includes blogs, homepages, and so on. Further, it is a concept including not only one page but also one constituted by a plurality of pages.

  In the embodiment, “character information acquisition means” corresponds to step S1 in FIG.

  In the embodiment, the “element information generation unit” corresponds to steps S2 to S5 in FIG.

  In the embodiment, “reception means” corresponds to step S11 in FIG.

  In the embodiment, “element information reading means” corresponds to step S12.

  In the embodiment, “search means” corresponds to steps S14 and S15.

  In the embodiment, “search result transmitting means” corresponds to step S16.

  The “program” is a concept including not only a program that can be directly executed by the CPU but also a program in a source format, a compressed program, an encrypted program, a program that can be installed and operated by a hard disk, and the like. .

BEST MODE FOR CARRYING OUT THE INVENTION

  FIG. 1 shows the configuration of a related web page search system according to an embodiment of the present invention. Web servers W1, W2 ... Wn are connected to the Internet 2. A terminal device 4 is connected to the Internet 2. In the figure, only one terminal device 4 is shown, but many terminal devices are actually connected. The terminal device 4 is generally a PC used by a user.

  A management server device 6 is further connected to the Internet 2. The management server device 6 includes character information acquisition means 8, element information generation means 10, reception means 12, element information reading means 14, search result transmission means 16, search means 18, and recording unit 20. The character information acquisition means 8 circulates each web page of the web servers W1 to Wn via the Internet 2 and acquires character information. In addition to the case where one web page is recorded on one web server, there are cases where two or more web pages are recorded on one web server. In this embodiment, even if there are a plurality of web pages, the web page created by the same person is assumed to be one web page even if it is composed of a plurality of pages.

  The element information generation means 10 analyzes the constituent elements of the character information for each acquired web page and generates web page element information. The generated web page element information is recorded in the recording unit 20 for each web page. At this time, the web page attributes such as the URL of the web page and the administrator (that is, the creator) are also recorded.

  The receiving means 12 receives a search request specifying the target web page from the terminal device 4. The element information reading unit 14 reads element information of the target web page indicated by the search request from the recording unit 20. The search unit 18 searches the recording unit 20 for another web page having element information similar to the element information of the read target web page. The search result transmission means 16 transmits attribute information such as the URL of the searched web page to the terminal device 4 that has made the request.

  As described above, a user who uses the terminal device 4 can know a web page related to a specific web page. For example, if a user's own web page is designated as a specific web page, a web page of another person similar to the user's own web page can be known.

  FIG. 2 shows a hardware configuration of the management server apparatus shown in FIG. A display 30, a communication circuit 34, a keyboard / mouse 36, a memory 38, a hard disk 40, and a CD-ROM drive 42 are connected to the CPU 32. The communication circuit 34 is a circuit for connecting to the Internet. In the hard disk 40, an operating system (OS) 46 such as WINDOWS (trademark), a management server program 48, a morpheme table 50, a circulation table 52, and the like are recorded. The management server program 48 functions in cooperation with the OS 46. These programs are those recorded on a recording medium such as the CD-ROM 44 and installed via the CD-ROM drive 42. It may be downloaded via the Internet.

  3 and 7 show flowcharts of the management server program recorded on the hard disk 40. FIG. FIG. 3 shows a process for creating a morpheme table, and FIG. 6 shows a process for searching a web page.

  In the following description, it is assumed that a person who operates a web server in order to cause many users to create a blog operates a management server apparatus 6. Therefore, the operator of the management server device 6 can easily know what kind of blog a user who is a member creates.

  The creation of the morpheme table in FIG. 3 is executed every predetermined period (for example, every day at 24:00). In step S1, the CPU 32 connects to the Internet via the communication circuit 34 and collects the description content (character information) of the web page. With reference to the circulation table 52 describing the top page URL of the circulation destination web page, connection is made to the web page.

  First, the CPU 32 acquires the URL of the web page described at the beginning of the circulation table 52 and connects to it. The CPU 32 acquires character information of the connected top page and records it in the memory 38. Further, a page created by the same creator as the creator of the top page is similarly acquired and recorded. In this embodiment, a link is pasted from the top page, and a page having a URL lower than the URL of the top page is created by the same person. For example, when the URL of the top page described in the circulation table is “http://www.furutani.co.jp/matsusihta”, the URL of the linked page is “http://www.furutani.co. If "jp / matsusihta / today.html", this page is determined to be a page created by the same person.

  When the CPU 32 acquires the character information of the web page in this way, the CPU 32 performs morpheme analysis on the character information with reference to a morpheme dictionary (not shown) (step S2). The CPU 32 performs morphological analysis and counts how many times each morpheme appears on the web page. In this embodiment, morphological analysis is performed with reference to a dictionary (not shown) in which morphemes are recorded. In this embodiment, morphological analysis is performed and only nouns are extracted.

  For example, when there is a web page as shown in FIG. 5, the number of appearances is counted for each morpheme such as patent, intellectual property, patent attorney. The CPU 32 calculates a score for each morpheme based on the counted number of appearances of each morpheme. In this embodiment, the score is calculated as follows.

Score = (number of appearances of the morpheme / total number of morphemes) * 100
In other words, the score is calculated depending on how much the morpheme appears with respect to the total number of morphemes on the web page. The higher the score, the more important the morpheme is on the web page. The CPU 32 records the score of each morpheme thus calculated in the morpheme table as shown in FIG. 6 (step S3).

  In this embodiment, since the creation of the morpheme table is executed once a day (24:00), the contents of the morpheme table are overwritten each time.

  Next, the CPU 32 determines whether or not the target web page is at the end of the circulation table (that is, whether or not all the web pages in the table have been visited) (step S4). If it is not the last web page, the CPU 32 sets the next web page in the list as the target web page (step S5), and repeatedly executes the processing from step S1.

  When the process for the last web page is performed in this way, the process for creating the morpheme table is terminated. An example of the created morpheme table is shown in FIG. In this embodiment, not only the score for each morpheme but also web page attributes such as title, URL, administrator (creator), etc. are recorded. For example, when the administrator of the website managing the blog creates a morpheme table for the web page of the blog creator who is a member, the attribute can be acquired based on the member information.

  Next, with reference to FIG. 7, a process when there is a search request from the user terminal device 4 will be described.

  A user who is a member first connects to the management server device 6 from the browser program of the terminal device 4. Further, upon receiving the search request screen as shown in FIG. 8, the search request is transmitted to the management server device 6. Here, it is assumed that the URL of one's blog is input and transmitted in the input box BOX1.

  The CPU 32 of the management server device 6 receives this search request through the communication circuit 34 (step S11). The CPU 32 specifies the target web page based on the URL included in the search request, and acquires the morpheme of the target web page from the morpheme table (step S12). At this time, the CPU 32 acquires a morpheme whose score exceeds “0” (that is, a morpheme that appears at least once). For example, if the target web page is “Akanebo” in FIG. 6, the morpheme “invention” is extracted, but the morpheme such as “patent” “invention” “patent attorney” is not extracted.

  Next, the CPU 32 extracts a predetermined number (five in this embodiment) of morphemes from the top of the score among the extracted morphemes (step S13). Subsequently, the CPU 32 extracts a web page including any of the five extracted morphemes (the score of the morpheme exceeds “0”) from the morpheme table (step S14). For example, if the five extracted morphemes are “patent”, “intellectual property”, “patent attorney”, “invention”, “development”, web pages such as “Intellectual Property Diary”, “Akanebo” and “Orange Pink” are selected. Will be.

  For each of the five morphemes, the CPU 32 extracts a predetermined number (three in this embodiment) of the selected web pages having a high score. That is, a total of 15 web pages are extracted, such as three web pages with high scores for the morpheme “patent”, three web pages with high scores for the morpheme “IP”, and the like. 38 (step S15).

  The CPU 32 converts the web page extracted for each morpheme into an image as shown in FIG. 9 and transmits it to the terminal device 4 that has made the request through the communication circuit 34 (step S16). In FIG. 9, the name of the target web page is displayed in the center, and five morphemes are shown around it. Furthermore, the name of the extracted web page is shown for each morpheme. The display of the name of each web page is a link to each web page. The CPU 32 can acquire this URL from the morpheme table and create this link.

  In this way, the terminal device 4 can know other people's blogs related to the morphemes that are key in their blogs. Since the morpheme table is updated every day, other similar blogs can be obtained at that time.

  In the above embodiment, the title of the web page is displayed in FIG. 9, but the creator of the blog may be displayed.

  In the above-described embodiment, a case has been described in which a web server that operates many users to create a blog operates a management server device 6. However, this system can be operated for web servers operated by others.

  In the above embodiment, in response to a request from the user, a search is performed and a related upper web page is transmitted. However, a search may be performed in advance, a search result may be recorded, and the recorded result may be transmitted when requested.

  In the above embodiment, the target web page is specified by the URL on the screen of FIG. However, it may be specified by the name of the web page, the e-mail address of the creator, or the like.

  In the above embodiment, the management server device 6 is realized by a single computer, but may be realized by a plurality of computers.

  In the above embodiment, the morpheme table is updated every day, and the old morpheme table is discarded (overwritten). However, the morpheme table on the first day of each month may be recorded in the hard disk 40 without being discarded. As a result, not only the current web page but also the past web page can be searched. For example, as shown in FIG. 10, the search result is displayed in such a manner that the time is specified in addition to the name of the web page. For example, “2002.5” in the “Patent Office Diary” is similar to the target web page when comprehensively judging the contents from the start of this web page to May 2002 It means that.

  Further, in the case of a diary or the like, when the creation period can be clearly grasped by the URL, a morpheme table may be generated for each period as shown in FIG. In the figure, the directory below http://www.furutani.co.jp/2003 indicates a diary described in 2003. If a morpheme is constructed in this way, each web page can be used as a search target for each period.

  Further, the search target can be limited to one web page, and each period can be set as the search target.

  In the above embodiment, a blog is shown as an example of a web page, but the same can be applied to a homepage or the like.

  In the above embodiment, the related web page is extracted for each of the morphemes extracted in step S13. Of the web pages including all five morphemes, a web page with a high score sum of each morpheme may be extracted. Further, the extracted web page may be used as a target web page, and a related web page may be further searched. As a result, as shown in FIG. 12, the related web page can be shown with a spread centering on the target web page “Intellectual Property Diary”.

  In the above embodiment, a web page similar to the user's own web page is searched, but a web page similar to another person's web page can also be searched. Further, on the input screen of FIG. 8, two web pages may be input, and the similarity between both web pages may be calculated. As the similarity, a numerical value obtained by adding up the scores of the morphemes having higher scores on both web pages and those on both web pages and dividing the sum by two can be used.

It is a functional block diagram of a related web page search system including a management server device according to an embodiment of the present invention. It is a figure which shows the hardware constitutions of a management server apparatus. It is a flowchart of a management server program. It is a figure which shows a circulation table. It is a figure which shows the example of a web page. It is a figure which shows the example of a morpheme table. It is a flowchart of a management server program. It is a figure which shows the example of a search screen. It is a display example of a search result. It is a display example of a search result. It is a figure which shows the example of the morpheme table by another example. It is a display example of a search result.

Explanation of symbols

8 ... Character information acquisition means 10 ... Element information generation means 12 ... Reception means 14 ... Element information reading means 16 ... Search result transmission means 18 ... Search means 20 ... Recording unit

Claims (11)

A management server device that searches other web pages highly relevant to a target web page on a network,
Character information acquisition means for accessing each web page and acquiring character information constituting each web page;
Element information generating means for extracting elements included in the acquired character information of each web page, generating web page element information, and recording it in the recording unit in association with the web page attribute information including the location specifying information of each web page When,
Receiving means for receiving a search request specifying the target web page sent from the terminal device;
Element information reading means for reading out the web page element information of the target web page specified by the search request from the recording unit;
Search means for referring to the recording unit and searching for a web page having web page element information similar to the read target web page element information to obtain web page attribute information;
Search result transmitting means for transmitting web page attribute information of the searched web page to the terminal device;
A management server device. A management server program for realizing, by a computer, the following means of a management server device for searching for another web page highly relevant to a target web page on a network:
Character information acquisition means for accessing each web page and acquiring character information constituting each web page;
Element information generating means for extracting elements included in the character information of each acquired web page, generating web page element information, and recording it in the recording unit in association with the web page attribute information including the location specifying information of each web page When,
Receiving means for receiving a search request specifying the target web page sent from the terminal device;
Element information reading means for reading out the web page element information of the target web page specified by the search request from the recording unit;
Search means for referring to the recording unit and searching for a web page having web page element information similar to the read target web page element information to obtain web page attribute information;
Search result transmission means for transmitting web page attribute information of the searched web page to the terminal device;
Server program for realizing the above with a computer. In the management server apparatus and program of Claim 1 or 2,
The element information generating means
Morpheme extraction means for extracting morpheme as an element by analyzing morpheme of character information on each web page
Score calculating means for calculating a score for each morpheme based on the appearance frequency of each extracted morpheme;
A morpheme table recording means for recording a score for each morpheme for each web page as a morpheme table in a recording unit;
With
The search means includes
Search morpheme extraction means for extracting a predetermined number of morphemes from the plurality of morphemes that are target web page element information based on the score,
With reference to the morpheme table, for each web page, a score given to the search morpheme is summed, and a web page selection unit that selects a similar web page based on the sum,
It is characterized by having. In the management server apparatus or program of Claim 3,
The score calculating means calculates the score of the morpheme based on the ratio of the number of applications of each morpheme extracted from the web page to the web page and the total number of morpheme on the web page. In the management server apparatus or program in any one of Claims 1-4,
The character information acquisition unit and the element information generation unit perform the operation every predetermined period. In the management server apparatus or program in any one of Claims 1-5,
Before receiving the search request, the element information reading unit and the search unit execute each web page search as a target web page and record it in the recording unit.
The search result transmitting means, when receiving a search request from a terminal device, transmits web page attribute information of a web page similar to the target web page recorded in the recording unit to the terminal device. In the management server apparatus or program in any one of Claims 1-6,
The search result transmitting means groups each web page based on a morpheme representing each searched web page and transmits it to the terminal device. In the management server apparatus or program in any one of Claims 1-7,
A second element information reading unit that reads the web page element information of the second target web page from the recording unit as a second target web page that is searched by the search unit;
A second search means for referring to the recording unit and searching for a web page having web page element information similar to the read second target web page element information;
With
The search result transmitting means transmits the web page attribute information of the web page found by both the search means and the second search means to the terminal device. In the management server apparatus or program in any one of Claims 1-8,
The element information generating means generates and records web page element information of each web page divided every predetermined period,
The search means specifies a period and searches for a related web page based on target web page element information for each predetermined period. In the management server apparatus or program in any one of Claims 1-9,
The web page is a blog. A web page search method for searching other web pages highly relevant to a target web page on a network by a computer,
The computer accesses each web page, obtains character information constituting each web page,
Extract the elements included in the character information of each acquired web page, generate web page element information, record it in the recording unit in association with the web page attribute information including the location specifying information of each web page,
Receives a search request sent from the terminal device that identifies the target web page,
Read the web page element information of the target web page specified by the search request from the recording unit,
Refer to the recording unit, search for a web page having web page element information similar to the read target web page element information, and obtain web page attribute information.
Transmitting the web page attribute information of the retrieved web page to the terminal device;
Web page search method characterized by

Images