JP2002132769A - Multidirectional search service method and apparatus, and recording medium recording the program - Google Patents

Abstract

(57) [Summary] [Problem] By examining search results from various viewpoints using search evaluation information other than the degree of coincidence with a keyword, a user can focus on a group of pages returned from a search engine. Make your pages easy to find. A search server is provided from user terminals.
When the search server 14 issues a request to re-sort the page information group with the evaluation information identifier and the reference page serving as the evaluation reference as arguments, the search server 14 issues an evaluation information acquisition request to the evaluation information server 15 and The server 15 calculates the evaluation information from each page in advance, registers the evaluation information in the evaluation information database 16, calculates the evaluation information when there is no evaluation information by searching, and charges the fee for the evaluation information. Returned to the search server 14, the search server 14 sorts the user terminals 11 to 1 in the order of the evaluation information.
Return to 3.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION The present invention relates to a search engine on the Internet that sorts search results using evaluation information other than the degree of matching with a keyword to examine the search results from various viewpoints. The present invention relates to a versatile search service method and apparatus, and a recording medium on which a program is recorded.

[0002]

2. Description of the Related Art Search engines use W on the Internet.
This is a system used when searching for a WW page (hereinafter simply referred to as a page). When a user inputs a keyword to a conventional search engine, the search engine returns a page including the keyword. However, when the number of keywords for search is reduced to one or a small number, the number of pages containing the keyword is often huge, and when the number of keywords is reduced to five or six, the number of pages is large. Many responded that there was no page that contained, and in the end it was necessary to find the best one among many pages.

Conventionally, in a search engine, as a technique for searching for a page of high importance from a plurality of pages as a search result, for example, Sergey Brin and
Lawrence Page: The anatom
y of a large-scale hypert
actual web search engine,
Proc. Of the Seventh Inter
national World Wide Web C
onference, pp. 107-117 (1998-4). Also, as a technology for searching for pages of high importance, Mo
nika Henzinger, Allan Heyd
on, Michael Mitzennmacher, a
nd Marc A. Najork. Measurin
g Index Quality using Ran
dom Walks on the Web. In P
rosedsing of the Eight I
international World Wide W
eb Conference, pages213-225 (1999-5).

Further, as a technique for searching for a page having high similarity from a plurality of pages as search results, for example, Ro
n Weiss, Bivenvenido Velez,
Mark A. Sheldon, Chanathip
Manprempre, Peter Sziergy
i, Andrzei David K .; Gifor
d. HyPursuit: A Hierarchica
l Network Search Engine
hat Exploits Content-Link
Hypertext Clustering. Pro
seedingsof the 7th ACM Co
nreference on Hypertext, New
There is York (1996).

[0005]

As described above, the conventional technique has a problem that it is extremely difficult for a user to find a desired page from a group of pages returned from a search engine on the Internet. there were. In other words, even if a search is requested using an appropriate keyword, there are many pages that match the keyword returned from the search engine, and it is necessary to search for the page desired by the user. The effort is more than imagined. In addition, in the above-mentioned foreign documents, a technology for searching for a high similarity from a plurality of searched pages or a technology for searching for a high importance is proposed. It was impossible to search for it in the end, and in the end, the user had to examine the search result from a multifaceted viewpoint.

Accordingly, an object of the present invention is to solve these conventional problems and to sort search results by using evaluation information other than the degree of matching with a keyword, thereby automatically obtaining search results from various viewpoints. It is an object of the present invention to provide a diversified search service method and apparatus capable of examining a program, and a recording medium recording the program.

[0007]

In order to achieve the above object, a multi-faceted search service method of the present invention provides a search server which issues a search request from a user terminal to a search server using a keyword as an argument. The page including the keyword is selected, the page information and the degree of matching with the keyword for each page information are calculated, the page information group is sorted in the order of the degree of matching, returned to the user terminal, and displayed on the user terminal. In the search service method of displaying the corresponding page selected by the user on the screen displayed, the returned user terminal sends the search information to the search server with the evaluation information identifier and the reference page as a reference for the evaluation as arguments. When you issue a request to re-sort information,
The search server relates to each page of the page information group,
The evaluation information server issues an evaluation information acquisition request to the evaluation information server using the evaluation information identifier and the reference page as arguments. The evaluation information server calculates evaluation information from each page on the Internet in advance, and outputs the evaluation information. Registered in the database, when receiving the evaluation information acquisition request, search the evaluation information database, when there is evaluation information on the page, obtain the evaluation information, return to the search server, When there is no evaluation information on the page, the evaluation information is calculated from the page, the calculation result is returned to the search server, and a fee is charged for the evaluation information. The information group is sorted in the order of the evaluation information and returned to the user terminal, and the user terminal displays the page information group. .

[0008] When the evaluation information identifier is the degree of coverage, the evaluation information server sets the target page to:
The number of links to the outside of the domain extended from the target page is obtained, and for each link to the inside of the domain extended from the target page, the number of extra-domain links extended from the linked page to The number obtained by multiplying the reciprocal of the number of links provided is added, and the sum of the two is returned to the user terminal via the search server as the coverage of the target page.

[0009] When the evaluation information identifier is an authority, the evaluation information server sends a message to the target page.
The number of links received from outside the domain set on the target page is calculated, and for each link from within the domain set on the target page, the number of linked links set outside the domain set on the link source page is added to the relevant page. Is added to the number multiplied by the reciprocal of the number of linked links, and the sum of the two is returned to the user terminal via the search server as the authority of the target page.

[0010] When the evaluation information identifier is the similarity (1), the evaluation information server sets an average of the average number of out-of-domain links for a group of pages in which the target page and the reference page are commonly linked out of the domain. The number obtained by dividing the natural logarithm of the value by the natural logarithm of the number of linked pages from the outside of the domain attached to the linked page is calculated, and the number added for the page group is calculated as the similarity between the target page and the reference page (1 ) And return it to the user terminal via the search server.

[0011] When the evaluation information identifier is the similarity (2), the evaluation information server determines the average number of domain outer link counts for a group of pages commonly linked from the target page and the reference page. The natural logarithm of the average value is divided by the natural logarithm of the number of links from the link source page to the outside of the domain, and the number obtained by adding the relevant page group is similarity between the target page and the reference page (2). Is returned to the user terminal via the search server.

When the evaluation information identifier is a graphic degree, a visual degree, or an audio degree, the evaluation information server determines the number of links from the target page to the image file, the moving image file, or the music file by using the audio information. It is returned to the user terminal via the search server as a degree, visual degree, or audio degree.

[0013] When the evaluation information identifier is of importance, the evaluation information server performs the following for all pages.
The constant a = 0.15 is divided by the total number of pages, and for the group of pages to which the target page links, the number obtained by dividing the importance of the linked page by the number of linked pages of the page is added for the group of pages. Multiplied by the constant b = 0.15, and for the group of pages linking to the target page, the number obtained by dividing the importance of the link source page by the number of links of the page is added for the group of pages The resulting number is multiplied by a number obtained by subtracting a constant a = 0.15 and a constant b = 0.15 from 1 and the above values are added to form a simultaneous equation. It solves by calculation, stores the result in a database, and returns the result as a degree of importance to the target page to the user terminal via the search server.

When the evaluation information identifier is freshness, the evaluation information server obtains a number obtained by subtracting the number obtained by subtracting the last update time from the current time by the update frequency from 1 to obtain the number. The number obtained by subtracting 1 from the number raised to the base e of the natural logarithm by a number is returned to the user terminal via the search server as the freshness.

When the evaluation information identifier is rich, the evaluation information server returns the file size of the target page as richness to the user terminal via the search server.

10) When the evaluation information identifier is a degree of popularity, the evaluation information server determines, for a group of users who clicked the target page in a unit time, the natural number of clicks of all users in the unit time. The logarithm is divided by the natural logarithm of the number clicked by the user in the unit time, and the number added for the user group is returned to the user terminal via the search server as popularity.

11) When the evaluation information identifier is the similarity (3), the evaluation information server calculates the average of the number of traffics with respect to the group of pages to which the user is commonly moving from the target page and the reference page. The search server divides the natural logarithm of the value by the number of users moving to the destination page, that is, the natural logarithm of the traffic number, and adds the number obtained for the page group as the similarity (3) between the target page and the reference page. Return to the user terminal via.

12) When the evaluation information identifier is the similarity (4), the evaluation information server sets the average value of the number of traffics for a group of pages to which the user is commonly moving from the target page and the reference page. The natural log of
The number of users moving from the source page, that is, the natural logarithm of the number of traffic, is divided by the natural logarithm, and the number added for the page group is used as the similarity (4) between the target page and the reference page via the search server. Return to terminal.

The diversified search service device of the present invention has 13)
When a request for re-sorting the page information group is received from the user terminal with the evaluation information identifier and the reference page serving as the evaluation reference as arguments, for each page of the page information group, the evaluation information identifier and the reference page are used as arguments. A means for issuing an evaluation information acquisition request for the page, an evaluation information database that calculates evaluation information from each page on the Internet in advance, and registers a calculation result, and the evaluation information identifier and the reference page as arguments, Means for retrieving the evaluation information database, means for obtaining evaluation information about the page, and returning the evaluation information; and when the evaluation information does not exist in the evaluation information database, the evaluation is performed from the page on the Internet. Means for calculating information and means for charging a fee for the evaluation information When, and means for sorting the page information group in the order of the evaluation information, the evaluation information and billing information and a means for returning to the user terminal.

14) Convert each processing step of the diversified search service method described in the above items (12) into a program and store the program in a recording medium.

[0021]

Embodiments of the present invention will be described below in detail with reference to the drawings. (Apparatus Configuration) FIG. 1 is a configuration diagram of a diversified search service apparatus according to an embodiment of the present invention. As shown in FIG. 1, user terminals 11, 12, and 13 used by each user are connected to the Internet 17, and a search server 14 provided by a search service provider and having a built-in search engine is connected. Is done. Up to this point, the configuration is the same as the conventional configuration, but in the present invention, in addition to this, an evaluation information server 15 provided by an evaluation information provider is newly connected to the Internet 17. An evaluation information database 16 is connected to the evaluation information server 15.

When an information evaluation request for a search result is issued to the search server 14 from the user terminals 11 to 13 from which a plurality of search results have been returned, the evaluation information server 15 sends an evaluation information acquisition request from the search server 14. It is activated by being issued. The evaluation information server 15 searches the evaluation information database 16 to obtain evaluation information on the page and returns it to the search server 14. In the evaluation information database 16, the evaluation information server 16 calculates and registers the evaluation information from each page on the Internet in advance. However, since the number of pages on the Internet increases every day, if the evaluation information has not been registered in the evaluation information database 16 for the target page for which the request has been issued, the evaluation information server 15 returns The evaluation information is calculated, and the calculation result is returned to the search server 14.

FIG. 20 is a detailed block diagram of the diversified search service device in FIG. User terminal device 20
Is a page moving unit 21 that performs an operation of moving by requesting a user from an arbitrary page to the next page.
After transmitting a search request to the search service providing apparatus (search server) 40 using the keyword as an argument,
A search service providing apparatus (search server) re-sorts a re-sort request using the reference page and the evaluation information identifier as arguments again.
The request result transmitting means 24 for transmitting the request result to the request 40, the request receiving means 22 for receiving the search result of the first search request and the request result of the re-sort request from the search service providing apparatus 40, And an operation execution unit 23 for executing the operation.

Search service providing device (search server) 40
Is a search means 44 for selecting a page including a keyword.
And a matching degree analyzing means 46 for calculating the matching degree of the selected page information, and an evaluation information providing device (evaluation information server) when a re-sort request is issued from the user terminal 20.
A re-sorting unit 45 for sorting the page information group from the server 30 in the order of the evaluation information, and a request receiving unit 41 for receiving a search request and a re-sorting request from the user terminal 20
And a page information group sorted in the order of the search page group and the evaluation information to the user terminal 20,
A request result transmitting unit 43 for issuing an evaluation information acquisition request to the evaluation information providing device 30 and an operation executing unit 42 for executing an operation necessary for the request are provided.

Evaluation information providing device (evaluation information server) 30
Is an evaluation information acquisition unit 32 that searches the evaluation information database device 36 to obtain evaluation information on the page.
Evaluation information charging means 31 for charging a fee for the evaluation information; request receiving means 33 for receiving an evaluation information acquisition request from the search service providing apparatus 40; and evaluation information and charging information for the search service providing apparatus 40. Request transmission means 35 for transmitting a request, and an operation execution means 34 for executing an operation necessary for these. The evaluation information provider that operates the evaluation information server 30 charges a fee for the evaluation information provided by the evaluation information charging unit 31. In this way, the evaluation information provider makes the search server 4
A business model that collects a fee for evaluation information from a search service provider that operates 0 can be constructed.

(Processing Operation) FIGS. 21 and 22 are diagrams showing examples of search results in the case of searching only for the degree of coincidence and in the case of searching with added authority. In FIG. 1, a user issues a search request to a search server 14 via the Internet 17 via user terminals 11 to 13.
In this case, the processing is the same as the conventional search method. At this time, the user specifies KW: fireworks as a parameter. The search service provider selects a page including the fireworks via the search server 14 and sets page information (title, summary,
URL) and the degree of coincidence. Title is "Fireworks"
The title and the outline of "Fireworks" are outlined in the contents of "Fireworks", UR
L is “protocol name / server name / file name”.
The WWW browser has a function of specifying a URL. The search service provider returns to the user terminals 11 to 13 the search result examples shown in FIG. In FIG. 21, URLs sorted in the order of 88, 77, and 66 are returned to the user terminal, starting with a page where 99 terms matching “fireworks” exist in the page.

Next, the user issues a re-sort request to the search server 14 via the user terminals 11 to 13. At this time, the user uses the evaluation information I as a parameter.
D: Specify authority. The search server 14 sends the evaluation information I to each URL of the search result.
D: A request for acquiring evaluation information relating to the URL is issued to the evaluation information server 15 using the authority as a parameter. The evaluation information server 15 stores the evaluation information database 1
6 to obtain evaluation information on the URL: authority. An example of the evaluation information shown in FIG. 22 can be obtained. The evaluation information server 15 returns the evaluation information example to the search server 14, and charges 1 yen for the evaluation information acquisition request. The search server 14 returns the search results sorted in the order of the evaluation information to the user terminals 11 to 13. The user terminal 11 to 13 displays the U
Go to RL and end the search service.

In FIG. 22, URLs sorted in the descending order of the authority values of the evaluation information are returned to the user terminal. A URL with a high authority is one in which the page is most frequently cited (linked) and is also linked to from the rest of the page, and in general, the parameters are used by ordinary users. Can be said to be a page that is always accessed.

(Example of Display Screen) FIG. 8 is an explanatory diagram showing an example of a screen when the present invention is executed. When executing the present invention, the screen shown in FIG. 8 is displayed on the display unit of the user terminal. First, the user inputs, for example, "fireworks" as a parameter to be searched in the input field, and clicks a search button on the right side.
100 items sorted in the order of 0 are searched. Next, when the user clicks the authority button among the authority, exhaustive, fresh, and abundant buttons arranged below, 100 items are provided which are sorted in the order of authority. An arrow below the authority button indicates that the first of the 100 cases has been displayed. Also, in order to specify the similarity, for example, after setting to the NTT homepage and clicking the button of “Similar page on the right”, a page similar to the NTT homepage is searched and output.

(Multiple Search Service Method) FIGS. 2 and 3 are processing flowcharts of the multilateral search service method according to one embodiment of the present invention. As shown in FIG. 2, in the process of the multifaceted search service method, first, a user issues a search request to a search server via a user terminal and a network, as in the related art. At this time,
The user specifies KW ("Fireworks" as an example) as a parameter (step 100). Next, the search service provider selects a page including the KW via the search server, and calculates page information (title, summary, URL) and the degree of matching. The search service provider returns the search results sorted in the order of matching degree to the user via the search server (step 101). The user terminal displays the search result and presents it to the user. The user can move to any page of the search result via the user terminal (step 102).

Next, the user issues a re-sort request via the user terminal to request the diversified search service of the present invention. At this time, the user specifies the evaluation information ID and the reference page (optional) as parameters. In response to the request, the search server sorts the search results by the specified evaluation information (step 103). The search is repeated for each URL on the search result, and when all the search is completed (step 104), as shown in FIG. 3, the search server returns the search results sorted in the order of the evaluation information to the user (step 107). ). Then, if it is finished (step 108), the processing is finished. If it is a new search (step 108), the process returns to step 100 again and repeats the processing up to this point. Otherwise,
Returning to step 102, the information is displayed on the screen of the user terminal and can be moved to an arbitrary page.

When the search is repeatedly performed for each URL in the search result of FIG. 2 (step 104), as shown in FIG. 3, the search server uses the evaluation information ID and the reference page (arbitrary) as parameters, and U
A request for obtaining evaluation information about the RL is issued (step 105). The evaluation information server searches the evaluation information database and acquires the evaluation information on the URL. If there is no evaluation information related to the URL in the evaluation information database, the user accesses a site on the Internet and calculates the evaluation information. The evaluation information server returns the evaluation information to the search server, and charges a fee for the evaluation information acquisition request. (For a detailed method of the evaluation information server and the evaluation information database, refer to Japanese Patent Application No. 2000-82111. (See specification and drawings) (step 106). Then, after repeatedly performing each URL on the search result (step 1
04) When the search is completed, the search results sorted in the order of the evaluation information are returned to the user as described above (step 107).

(Evaluation Information Calculation Method) FIGS. 4 to 7 are processing flowcharts of an evaluation information calculation method in the evaluation information server of the present invention. FIGS. 9 to 19 show evaluation information showing one embodiment of the present invention. It is explanatory drawing which shows the calculation formula. If the evaluation information server receives the evaluation information acquisition request (step 111), the evaluation information ID shown in FIG.
The number of links to the outside of the domain extended from the page x with respect to the page x according to the formula shown in FIG. The number obtained by adding the number obtained by multiplying the number of links by the reciprocal of the number of links stretched from page y is calculated as the coverage of page x, and the result is returned (step 112). here,
The number multiplied by the reciprocal is a weight, for example, page y
Assuming that the number of links outside the domain spanned from page 3 is 2 and the number of links outside the domain spanned from page x is 2, the reciprocal (weight) of only the former is added, and (1/3) × 3 + 2 = 3
Is a value that contributes as a whole. In other words, a page with a high degree of coverage has many links to other pages and also has many links within the same page.

Next, when the evaluation information ID is the authority, FIG.
For the page x, the number of links from outside the domain attached to the page x, and for each link from within the domain attached to the page x, the outside of the domain attached to the link source page y Calculate the sum of the number of linked pages multiplied by the reciprocal of the number of linked pages attached to page y, and add the number as the authority of page x, and return the calculation result (step 113). For example, if the number of links to page x is 2, page y to page x
If the number of out-of-domain links is 3, the reciprocal (weight) of only the latter is added, and (１ ／) × 3 + 2 = 3 is a value that contributes as a whole. In other words, a page with a high authority is a page that is frequently referred to from other pages and is also referred to frequently from within the same page. The difference between the expression of coverage and the expression of authority is that the expression of coverage is linked in the direction of x → y, whereas the expression of authority is linked in the direction of y → x. Are different

Next, when the evaluation information ID is the similarity (1),
According to the formula shown in FIG. 11, page x and reference page y
Regarding pages that are commonly linked outside the domain,
The number obtained by adding the natural logarithm of the average value of the average number of links outside the domain (eg, 10) by the natural logarithm of the number of links from outside the domain attached to the linked page A is added to the page x. And the similarity (1) of the reference page y and return the calculation result (step 114). That is,
A page having a high similarity (1) is a page having a large number of links to outside the domain in common with the reference page. In the formula of FIG. 11, the average number of out-of-domain links is normalized by a value (fraction) obtained by dividing the number of in-links extending from the outside of the domain to the page x, thereby facilitating use. In addition, since the average value of the average number of links outside the domain has been optimally set to 10, conventionally, the present embodiment also follows this.

Next, when the evaluation information ID is the similarity (2),
According to the formula shown in FIG. 12, page x and reference page y
, The natural log of the average value of the average number of domain coat links (eg, 10) for the group of pages that are commonly linked outside the domain is calculated as the natural log of the number of links from the link source page A to the outside of the domain. Is calculated as the similarity (2) between page x and reference page y, and the calculation result is returned (step 115). The difference between the similarities (1) and (2) is that the former is the reference page y
And page x are commonly linked to external page a, whereas the latter is that reference page y and page x are commonly linked from external page a. In other words, pages having a large similarity (2) are similar to each other as the number of pages linked from the same page increases.

Next, when the evaluation information ID is a graphic degree, a visual degree, and an audio degree, the number of links from the page x to the image file, the moving image file, and the music file is calculated according to the formula shown in FIG. , Returned as audio degree. An extension is used as a criterion for determining a link to each file (step 116).
In other words, a page having a large graphic degree, visual degree, and audio degree corresponds to a page having a large number of links to image files, moving image files, and music files.

Next, when the evaluation information ID is of importance, FIG.
According to the formula shown in FIG.
(Example: 0.15) divided by total page T and constant b
(Example: 0.15) is multiplied by the number obtained by adding the number obtained by dividing the importance of the linked page p by the number of linked pages of the page p with respect to the page group to which the page x links. The number obtained by subtracting the constants a and b from 1 and the number obtained by dividing the importance of the link source page q by the number of links of the page q for the page group linking the page x is added. Create an expression that multiplies a number by a number. The multidimensional simultaneous equations created by the above procedure are solved by numerical calculation, and the results are stored in a database. At the time of the request, the importance for page x is returned (step 117). In the equation shown in FIG. 14, a constant a = 0.15 is known as an optimum value in the conventional index, and therefore this value is also used in the present embodiment. Also, the constant b
Although the optimum value has not been determined yet, b = 0.15 is used as the optimum value in this embodiment. Further, in the conventional expression indicating importance, it is usual that there is no last term of the expression shown in FIG. 14, but by using the expression in which the last term is added in this embodiment, a more accurate solution can be obtained. It turned out to be obtained.

The meaning of the equation shown in FIG. 14 indicates the probability that the user is viewing page x of all pages (for example, 1 million pages) in the term a · 1 / T, and the following Linki
In the term of ng (x), when observed at a certain point in time, the probability that the page x is seen is a part indicating how many out of 1 million people are. When one of them is from page x, 1/10 of the importance of p is the contribution from page x. Next Link
The term ed (x) is a value obtained by multiplying the value obtained by dividing the importance of the link source page q by the number of linked pages placed on the page q by subtracting a and b from 1. This is the new addition. For example, if the simultaneous equations are 100
Since only all expressions can be described, the algebraic expressions are solved and the calculated values become the importance.

Next, when the evaluation information ID is the freshness, FIG.
The number obtained by subtracting the number obtained by subtracting the last update time from the current time by the update frequency is subtracted from 1 in accordance with the equation shown below, and the number obtained by subtracting 1 from the number obtained by raising the base e of the natural logarithm is taken as the freshness. Calculate and return the calculation result. For example, the last update time is obtained from the Last-modified tag of the HTTP header, and the update frequency is obtained by regular monitoring (step 118).

Next, when the evaluation information ID is rich, FIG.
The file size of page x is returned as abundance according to the formula shown in (1) (Step 119). For example, the file size is the Content-Length of the HTTP header.
Get from the th tag.

Next, when the evaluation information ID is popularity, FIG.
The natural logarithm of the average number of clicks (eg, 1 click / hour) of all users within the unit time for the group of users who clicked the page x within the unit time (eg, 1 hour) is calculated according to the formula shown in Then, the number calculated by adding the number of user clicks divided by the natural logarithm is calculated as popularity, and the calculation result is returned (step 120). In this way, the expression of popularity is normalized. This is 10 per day
This is to increase the contribution of the latter who does not treat the case where a person who clicks on 100,000 pages sees that page and the case where only 10 pages a day clicks on that page look at the same page. .

Next, when the evaluation information ID is the similarity (3),
According to the formula shown in FIG. 18, page x and reference page y
From the group of pages to which the user is commonly moving, the natural logarithm of the average value of the number of trafficed is divided by the natural logarithm of the number of users (traffic number) moving to the destination page A, and the sum is added. Is calculated as the similarity (3) between page x and reference page y, and the result is returned (step 121). In the above similarity (1) and similarity (2), the number of direct links was calculated.
In the expression of the similarity (3) shown in FIG. 18, if the number of users moving (traversing) from the page x and the reference page y to another page A is seen, the greater the number, the more the similarity (3) Is high.

Next, when the evaluation information ID is similarity (4),
According to the formula shown in FIG. 19, page x and reference page y
For a group of pages to which the user is commonly moving, the natural logarithm of the average value of the number of traffic is divided by the natural logarithm of the number of users (the number of traffic) moving from the source page A and added. Number is page x and reference page y
Is calculated as (4), and the calculation result is returned (step 122). The above-mentioned similarity (3) calculates the number of people who move from the page x and the reference page y to another page A. The similarity (4) shown by the equation in FIG. To calculate how many people are moving in common to pages x and y.

Each processing step of the diversified search service method shown in FIGS. 2 to 3 and the evaluation information calculation method shown in FIGS. 4 to 7 is converted into a program, and the converted program is
By storing the recording medium in a recording medium such as a D-ROM, the recording medium can be mounted on any computer connected to the Internet, and the program can be installed or downloaded to another computer via the Internet. The present invention can be easily realized by those computers.

[0046]

As described above, according to the present invention,
Coverage, authority, similarity (1) (2), graphicalness, audioness, visuality, importance, freshness, popularity, similarity (3) (4), and similarity with keywords By sorting search results using search evaluation information other than, search results can be examined from multiple perspectives, so users can easily find the page they want from the group of pages returned from the search engine. It will be possible to discover.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of a diversified search service device according to an embodiment of the present invention.

FIG. 2 is a flowchart (1) of a multi-faceted search service method according to an embodiment of the present invention.

FIG. 3 is a process flowchart (2) of the multifaceted search service method.

FIG. 4 is a processing flowchart of an evaluation information calculation method (coverage, authority, similarity (1)) showing one embodiment of the present invention.

FIG. 5 is a processing flowchart of an evaluation information calculation method (similarity (2), graphic level, visual level, audio level, importance level) showing an embodiment of the present invention.

FIG. 6 is a processing flowchart of an evaluation information calculation method (freshness, abundance, popularity) showing an embodiment of the present invention.

FIG. 7 is a processing flowchart of an evaluation information calculation method (similarity (3), similarity (4)) showing one embodiment of the present invention.

FIG. 8 is a diagram of a multifaceted search service screen showing one embodiment of the present invention.

FIG. 9 is an explanatory diagram showing a calculation formula of the coverage in FIG. 4;

FIG. 10 is an explanatory diagram showing a formula for calculating the authority in FIG. 4;

FIG. 11 is an explanatory diagram showing a formula for calculating the similarity (1) in FIG. 4;

FIG. 12 is an explanatory diagram showing a calculation formula of a similarity (2) in FIG. 5;

FIG. 13 is an explanatory diagram showing a calculation formula of a graphical degree, a visual degree, and an audio degree in FIG. 5;

FIG. 14 is an explanatory diagram showing a formula for calculating importance in FIG. 5;

FIG. 15 is an explanatory diagram showing a calculation formula of freshness in FIG. 6;

FIG. 16 is an explanatory diagram showing a formula for calculating abundance applied to FIG. 6;

FIG. 17 is an explanatory diagram showing a formula for calculating the degree of popularity in FIG. 6;

FIG. 18 is an explanatory diagram showing a calculation formula of the similarity (3) in FIG. 7;

FIG. 19 is an explanatory diagram showing a formula for calculating the similarity (4) in FIG. 7;

FIG. 20 is a detailed configuration diagram of the diversified search service device of FIG. 1;

FIG. 21 is a diagram showing an example of a conventional search result sorted.

FIG. 22 is a diagram showing an example of a sorted search result according to the present invention.

[Explanation of symbols]

11 to 13: user terminal, 14: search server, 15: evaluation information server, 16: evaluation information database, 17, 5
0: Internet, 21: Page moving means, 22, 3
3, 41 ... request receiving means, 23, 34, 42 ... operation executing means, 24, 35, 43 ... request result transmitting means, 32 ... evaluation information acquiring means, 31 ... evaluation information charging means, 36 ... evaluation information database device, 44: searching means, 45: re-sorting means, 46: matching degree analyzing means, 20: user terminal device, 30: evaluation information providing device, 40: search service providing device.

Claims (14)

[Claims] 1. When a search request is issued from a user terminal to a search server using a keyword as an argument, the search server selects a page including the keyword, and matches the page information with the keyword for each page information. In the search service method that calculates the degree, sorts the relevant page information group in the order of coincidence and returns it to the user terminal, and displays the corresponding page selected by the user on the screen displayed on the user terminal. When the user terminal issues a request for re-sorting the page information group with the evaluation information identifier and the reference page serving as an evaluation reference as arguments to the search server from the user terminal, the search server sets each of the page information groups. For a page, the evaluation information identifier and the reference page are arguments to the evaluation information server, and the evaluation information for the page is The evaluation information server calculates the evaluation information from each page on the Internet in advance and registers the evaluation information in the evaluation information database. When the evaluation information acquisition request is received, the evaluation information server reads the evaluation information database. When the search is performed and the evaluation information on the page is present, the evaluation information is acquired and returned to the search server. When the evaluation information on the page is not present, the evaluation information is calculated from the page, and the calculation result is obtained. Is returned to the search server, and a fee is charged for the evaluation information. The search server sorts the received page information group in the order of the evaluation information and returns the page information group to the user terminal. A multifaceted search service method characterized by displaying a group of page information. 2. The multi-faceted search service method according to claim 1, wherein when the evaluation information identifier is a coverage degree, the evaluation information server sets a domain set from the target page to the target page. The number of outgoing links is determined, and for each link in the domain spanned from the target page, the number of out-of-domain links spanned from the linked page is
Multiplying the number obtained by multiplying the reciprocal of the number of links provided from the page, and returning the sum to the user terminal via the search server as the coverage of the target page. Search service method. 3. The multi-faceted search service method according to claim 1, wherein, when the evaluation information identifier is an authority, the evaluation information server sets a domain assigned to the target page for the target page. Obtain the number of inbound links from outside, and for each link from within the domain attached to the target page, add
A polygon characterized by adding the number obtained by multiplying the reciprocal of the number of linked pages attached to the page, and returning the added number as the authority of the target page to the user terminal via the search server. Search service method. 4. The multi-faceted search service method according to claim 1, wherein when the evaluation information identifier is the similarity (1), the evaluation information server links the target page and the reference page out of the domain in common. For the group of pages, calculate the number obtained by dividing the natural logarithm of the average value of the average number of links outside the domain by the natural logarithm of the number of links from outside the domain attached to the linked page, and add it for the page group A multi-faceted search service method characterized by returning the determined number as the similarity (1) between the target page and the reference page to the user terminal via the search server. 5. The multi-faceted search service method according to claim 1, wherein when the evaluation information identifier is the similarity (2), the evaluation information server is commonly linked outside the domain from the target page and the reference page. For the group of pages, the natural logarithm of the average value of the average number of domain envelope links was divided by the natural logarithm of the number of links from the link source page to the outside of the domain, and added to the group of pages A multi-faceted search service method characterized by returning the number as a similarity (2) between the target page and the reference page to the user terminal via the search server. 6. The multi-faceted search service method according to claim 1, wherein the evaluation information identifier is a graphic degree, a visual degree,
Or, when the audio level is used, the evaluation information server uses the number of links from the target page to the image file, the video file, or the music file as the audio level, the visual level, or the audio level, respectively, via the search server. A multi-faceted search service method characterized by returning to a user terminal. 7. The multi-faceted search service method according to claim 1, wherein when the evaluation information identifier is the importance level, the evaluation information server sets a constant a = 0.15 for all pages. For a group of pages to which the target page links, the constant b = 0.15 is added to the number obtained by dividing the importance of the linked page by the number of linked pages of the page and adding the number of the linked pages to the group of pages. And the number obtained by adding the number obtained by dividing the importance of the link source page by the number of links of the relevant page to the page group linking the target page and adding the number of the relevant page group from 1 to a constant a = 0 .1
5. Multiply by the number obtained by subtracting the constant b = 0.15, add the above values to create a simultaneous equation, solve the created multidimensional simultaneous equation by numerical calculation, and store the result in the database And returning the result as a degree of importance to the target page to the user terminal via the search server. 8. The multi-faceted search service method according to claim 1, wherein when the evaluation information identifier is freshness, the evaluation information server divides the number obtained by subtracting the last update time from the current time by the update frequency. The number obtained by subtracting 1 from 1 is obtained, and the number obtained by subtracting 1 from the number obtained by raising the base e of the natural logarithm to the number is returned as a freshness to the user terminal via the search server. Diversified search service method. 9. The multi-faceted search service method according to claim 1, wherein when the evaluation information identifier is rich, the evaluation information server sets the file size of the target page as rich and sends it via the search server. A multifaceted search service method characterized by returning to a user terminal. 10. The multi-faceted search service method according to claim 1, wherein when the evaluation information identifier is a degree of popularity, the evaluation information server determines whether a user group clicked a target page within a unit time for a unit time. The natural log of the average number of clicks by all users within the unit is divided by the natural log of the number of clicks by the user in a unit time, and the number added for the group of users is used as the degree of popularity, and the number of users is calculated via the search server. A multifaceted search service method characterized by returning to a terminal. 11. The multi-faceted search service method according to claim 1, wherein when the evaluation information identifier is the similarity (3), the user moves from the target page and the reference page in common to the evaluation information server. For the group of pages, the natural logarithm of the average value of the number of trafficees is divided by the number of users moving to the destination page, that is, the natural logarithm of the number of traffic, and the number added for the page group is used as the target page A multi-faceted search service method characterized by returning a page similarity (3) to a user terminal via a search server. 12. The multi-faceted search service method according to claim 1, wherein when the evaluation information identifier is the similarity (4), the evaluation information server moves the user in common from the target page and the reference page. For a group of pages, the natural log of the average value of the number of traffic is divided by the number of users moving from the source page, that is, the natural log of the number of traffic, and the number added for the page group is used as the target page as a reference. A multi-faceted search service method characterized by returning a page similarity (4) to a user terminal via a search server. 13. A multi-faceted search service apparatus for searching for a page on the Internet, wherein a re-sort request for the page information group is received from a user terminal using an evaluation information identifier and a reference page as a reference for evaluation as arguments. Sometimes, for each page of the page information group, means for issuing an evaluation information acquisition request for the page using the evaluation information identifier and the reference page as arguments, and calculating the evaluation information in advance from each page on the Internet, and calculating the calculation result. A means for searching the evaluation information database using the evaluation information identifier and the reference page as arguments, a means for obtaining evaluation information about the page and returning the evaluation information, When the evaluation information does not exist in the evaluation information database, Means for calculating the evaluation information from the page, means for charging a fee for the evaluation information, means for sorting the page information group in the order of the evaluation information, and storing the evaluation information and the charging information in the user terminal. And a return means. 14. A computer characterized in that each processing step of the diversified search service method according to any one of claims 1 to 12 is converted into a program, and the program is stored in a recording medium. A readable program recording medium.