JP2006338346A - Retrieval server - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a retrieval server for enabling a user to quickly reach his/her desired site or contents on the Internet with a high possibility. <P>SOLUTION: The retrieval server 1 stores meta data 22 and contents evaluation values 23 of a plurality of sites in a contents DB 21. When the user makes a retrieval request from the top page 26 of a retrieval site 2, the existence/nonexistence of corresponding contents 25 including a retrieval word 24 is retrieved in the meta data 22, the evaluation calculation value of each site is calculated with the evaluation values 23 of contents as a base in the site, and the total evaluation calculation value is calculated from the evaluation values of the entire contents of the site. Each site is sorted with the evaluation calculation value as a base, and a sorted site is displayed on a retrieval result screen 28. Since the number 38 of corresponding contents, evaluation calculation values 39, the total number 40 of contents and the total evaluation calculation value 41 are displayed at that time, the user can recognize evaluations of the site and contents. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a search server used when searching a website or the like on a network such as the Internet.

  Conventionally, many sites exist on a network such as the Internet. For this reason, a user searches for a site having information such as content desired by the user using a search server called a so-called search engine.

  In such a search site, a user performs a search by inputting a search term into the search term input field of the site. The search results are generally extracted from sites that contain the search term in the site data (site name, description, web page, text content, etc.) and list the extracted sites. It is.

  However, in the conventional search site, if the search term data is included in the data of the search target site, for example, the list is often displayed in descending order of the number of search terms. This means that there are many search terms on the site, and no consideration is given to the quality of the site, so the site desired by the user is ranked higher just because it is ranked higher as a search result. It is not always done. In addition, since there are many simple descriptions of sites displayed in the search results, it is difficult to determine the quality of the sites based only on the descriptions. For the above reasons, it is very difficult to find the content or site desired by the user when the number of extracted sites is large.

  On the other hand, Patent Document 1 below discloses a search engine that can perform a search by designating content desired by a user. For example, when the user desires a certain kind of image, a name (search word) related to the image is input to the search word input field and the search target is limited to the image. Then, first, as in the case of a conventional search engine, a site including a search term is extracted, content having an extension related to an image (such as .gif or .jpg) in the site is extracted, and the corresponding result is displayed on the search result screen. Content with an extension related to images is displayed as a hyperlink in association with the site. As described above, in the search engine disclosed in Patent Document 1, the hyperlink of the content desired by the user is added to the search result of the normal search engine and displayed, and the user clicks the hyperlink. You can link directly to the content you want.

However, in the search engine disclosed in Patent Document 1, the search result of the site is not different from that of the conventional search engine. Therefore, the site that the user really wants is not necessarily listed at the top. For this reason, even if a link is directly provided to the content of the site, the probability that the content is desired by the user is low. Further, although a hyperlink of content is displayed in the search result list, since only the file name of the content indicates the content, the content of the content is often unknown. Furthermore, since hyperlinks of content are displayed in the search result list, if there are many contents on the extracted site, many hyperlinks are displayed on the search result display screen. There is also the inconvenience that it becomes difficult to see.
JP 2000-76272 A (paragraph number 0026,0027, FIG. 9)

  The present invention aims to improve a search server used when searching a website or the like on a network such as the Internet, and more specifically, in order to eliminate the inconvenience, a user can quickly reach a site or content desired by the user. An object is to provide a search server that is highly possible.

  In order to achieve the above object, the search server according to the first aspect of the present invention includes, for a plurality of contents existing in a plurality of sites, at least an evaluation value of the contents and a URI of the site where the contents exist. When a search by a search word is performed via a network from a content database in which metadata having information on the content is stored, corresponding content including the search word in the metadata is extracted from the content database Extraction means, evaluation calculation value calculation means for calculating evaluation calculation values of all corresponding contents for each site based on an evaluation value of the corresponding contents, and the number of the evaluation calculation values or the corresponding contents Sorting means to sort each site on the basis of, and sorted For site, characterized in that a display means for browsable displayed to the user with the number of the evaluation calculation value or the corresponding content.

  The search server according to the first aspect of the present invention is a system that calculates and sorts the evaluation calculation values of sites extracted by a search when a search is performed by a user using a search word. Specifically, the metadata of the content is stored in the content database, and when the user performs a search using a search term, the extraction unit searches for the metadata. Then, contents including search terms in the metadata are collected for each site, and the evaluation calculation values of all corresponding contents are calculated by the evaluation calculation value calculation means. Further, the sorting unit sorts the sites based on the evaluation calculation value or the number of the corresponding contents. Then, the evaluation means or the number of corresponding contents is displayed to the user in the sorted site order by the display means.

  Therefore, according to the search server of the 1st mode of the present invention, a user can peruse a site which has contents about a search word in order of a site where contents with high evaluation exist. At that time, since the evaluation calculation value and the number of the corresponding contents are displayed, the user can search how many contents the user desires in the site and what the evaluation of the contents is. You can check on the screen. Thus, the user can quickly reach the desired site from the search result screen without browsing each site directly from the search result screen. Note that only one of the evaluation calculation value and the number of the corresponding contents may be displayed, or both may be displayed.

  Here, URI (Uniform Resource Identifiers) indicates a framework for indicating location information including http, http, or urn (Uniform Resource Name) used on the Internet. Uniform Resource Locator) is known.

  In the search server of the first aspect of the present invention, the evaluation calculation value calculation means calculates a total evaluation calculation value by a second calculation formula based on the evaluation of all contents in each site, and the display means The total evaluation calculation value and the total number of contents on the site may be further displayed. According to this, not only the number of content and the evaluation value that the user desires, but also how much content exists on the site where the content exists, and the evaluation value of the content existing on the site is I can know. For this reason, it is possible to refer to the scale and evaluation of the site where the content desired by the user exists, so that the value of each site can be judged together with the number and evaluation of the content itself. Discovery is easier. This second calculation formula may be the same as the calculation formula used by the evaluation calculation value calculation means, or may be another formula.

  Further, the search server according to the second aspect of the present invention provides metadata having information on the content including at least the evaluation value of the content and the URI of the site where the content exists, for a plurality of content existing in the plurality of sites. A content database in which is stored, first extraction means for extracting corresponding content including the predicted search word in the metadata from the content database using a predicted search word expected to be searched by a user, and the corresponding Evaluation calculation value calculation means for calculating evaluation calculation values of all corresponding contents for each site by a predetermined calculation formula based on the evaluation value of the contents, and each site based on the evaluation calculation value or the number of the corresponding contents A sorting means for performing sorting, a sorted result, the evaluation calculation value, and the The result storage means for storing the number of contents, and when the user performs a search using the actual search word via the network, the predicted search word corresponding to the actual search word is extracted from the result storage means And a display means for displaying the sort result site corresponding to the predicted search word extracted by the second extraction means so that the user can view the information together with the evaluation calculation value or the number of the corresponding contents. It is characterized by having.

  The search server according to the second aspect of the present invention predicts a search word used by a user in advance, extracts the corresponding content in advance using the predicted search word, and extracts the corresponding content for each site extracted by the search. The calculation of the evaluation calculation value and the sorting by the sorting means are performed, and when the search is actually performed by the search word from the user, the search result of the expected search word corresponding to the actual search word is displayed. It is.

  Therefore, similarly to the first aspect of the present invention, the user can browse the sites in the order of the sites where the highly evaluated content exists for the sites having the content related to the search term. Furthermore, when the user actually performs a search, the search using the predicted search term has been completed in advance, and only the search result is displayed to the user. Even so, it is possible to promptly display search results and the like to the user, and to reduce the load on the search server. Here, the predicted search word includes not only one word but also a combination of words.

  In the search server according to the second aspect of the present invention, when there is no predicted search word corresponding to the actual search word, the first extraction unit searches the metadata from the content database for the actual search. The content including words is extracted, the evaluation calculation value calculation means calculates the evaluation calculation values of all the corresponding content for each site by a predetermined calculation formula based on the evaluation value of the corresponding content, and the sorting means Each site is sorted based on the evaluation calculation value or the number of the corresponding contents, and the display means displays the sorted sites together with the evaluation calculation value or the number of the corresponding contents so that the user can view it. Is preferred. As described above, when there is no predicted search word corresponding to the actual search word, the user can perform the processing by each means when the search is performed by the user. Similar search results can be displayed.

  In the search server of the second aspect of the present invention, the evaluation calculation value calculation means calculates a total evaluation calculation value by a second calculation formula based on the evaluation of all contents at each site, and stores the result storage The means may further store the total evaluation calculation value and the total number of contents in each site, and the display means may further display the total evaluation calculation value and the total number of contents in the site. According to this, as in the first aspect, the user can determine not only the number and evaluation value of the content he desires, but also how much content exists on the site where the content exists, and the site. It is possible to know the evaluation value of the content existing in.

  In the first or second aspect of the present invention, the display means may display the evaluation calculation value and the number of corresponding contents as an image, and the total evaluation calculation value and content in each site. The total number of images may be displayed as an image. The evaluation calculation value and the number of the corresponding contents, or the total number of contents existing on the site and the total evaluation calculation value thereof may be expressed by numbers, but may be expressed by numbers, but when displayed as an image such as a bar graph Evaluation calculation values and these numbers can be grasped intuitively by the user.

  In the first or second aspect of the present invention, the display means displays the evaluation calculation value, the total evaluation calculation value, the content evaluation value, the number of corresponding contents, or the total number of contents in color. Also good. For example, in the numerical value or image such as the evaluation calculation value, the high evaluation is expressed in red, and the user is more intuitive by attaching colors such as orange, yellow, green, and blue as the evaluation becomes low. It becomes possible to select content or a site having a high evaluation.

  In the first or second aspect of the present invention, the display means may further display the type of content. The content type is, for example, “jpg” or “gif” for an image, and “rm” or “mov” for a moving image. By indicating the type of content in this way, the type of content of the site extracted by the search can be confirmed at a glance, so that the user can further easily select the content or site.

  Next, an embodiment of a search server of the present invention will be described with reference to FIGS. First, a first embodiment of the present invention will be described with reference to FIGS. In the first embodiment, when a user performs a site search using a search word, a search is actually performed and a search result is displayed to the user. As shown in FIG. 1, the search server 1 of the first embodiment performs a search in a search site 2 called a so-called search engine. As shown in FIG. 1, the search server 1 is connected to the Internet 3 (network), and the Internet 3 registers the search server 1, the user terminal 5 used by the content user 4, and the content. A registrant terminal 7 used by the registrant 6 is connected to a web server 8 on which a home page (web site) of the registrant 6 including content is registered.

  A search site 2 shown in FIG. 1 is operated by an operator of the search server 1. In addition, the web server 8 is operated by a company separate from the search server 1, and in this embodiment is a web server of a network connection company, and a general user can open a homepage for free or for a fee. ing. In addition, on the homepage 9 of the registrant 6, there are a total of nine photos including five panda photos and four other photos as content.

  As shown in FIG. 2, the main functional configuration of the search server 1 includes hardware such as a CPU 10, a memory 11, a hard disk 12, and a router 13, and software stored in the hard disk 12. As software, an extraction program 14 (extraction means), an evaluation calculation value calculation program 15 (evaluation calculation value calculation means), a total evaluation calculation value calculation program 16 (evaluation calculation value calculation means), a sort program 17 (sort means), In addition, a display program 18 (display means) is stored. A content evaluation program 19 for evaluating the content itself and a check program 20 for checking the validity of the search request to the search site 2 are also stored in the hard disk 12. The storage area of the hard disk 12 stores content metadata 22 entered on a registration screen 27 and the like, which will be described later, and a content evaluation value 23 as a content database 21 (hereinafter referred to as a content DB). Yes.

  The extraction program 14 searches whether or not the search word 24 is included in the metadata 22 of the content registered in the content DB 21 when the search by the search word 24 is performed by the user 4. The content 24 is included as the metadata 22, that is, the corresponding content 25 is extracted. The evaluation calculation value calculation program 15 summarizes the corresponding content 25 belonging to the same site with respect to the corresponding content 25 extracted by the extraction program 14, and calculates the evaluation value of the corresponding content 25 for each site by a predetermined calculation method. The value is calculated. The total evaluation calculation value calculation program 16 calculates the evaluation values of all contents (corresponding contents and other contents) registered in the site at a site having the corresponding content 25 by a predetermined calculation method (second calculation formula). ) To calculate a total evaluation value that is an evaluation value of all the contents of the site.

  The sort program 17 rearranges the sites in the descending order of evaluation calculation values for the sites extracted by the extraction program 14. The display program 18 displays the sites sorted by the sort program 17 on the user terminal 5 through a client program such as a web browser or an RSS reader together with evaluation calculation values and the like.

  The content evaluation program 19 evaluates a single content based on various information such as the number of accesses to the content, the viewer, the remote host name of the user, the IP address, or the reference source URL. Specifically, the number of accesses is counted based on the number of download requests and browsing requests for the content. In addition, the remote host name or the like can be recognized by detecting a duplicate host (user terminal) by the remote host name and excluding points. For example, it is possible to exclude from the evaluation an unjust act that causes the content to be displayed many times a day in order to improve the evaluation of its own content.

  The metadata 22 stored in the content DB 21 is the content registered on a registration screen 27 described later. The content evaluation value 23 is a value evaluated for each content by the content evaluation program 19.

  Next, various screens displayed on the user terminal 5 in the search server 1 of the first embodiment will be described. On the user terminal 5, a top page 26, a registration screen 27, and a search result screen 28 of the search site 2 are displayed. These displays are displayed by the display program 18 via a web browser built in the user terminal 5.

  As shown in FIG. 3, the top page 26 is used when the user 4 registers contents in the search site 2 in addition to the input area 29 for entering the search word 24 and the search button 30 for executing the search. A new registration button 31 used for the above is provided. In addition, a link (characters having an underline in the figure) to a search page for each content (not shown) for searching for each content type is also provided. In the present embodiment, links are provided that move to search pages of the image 32, the moving image 33, and the blog 34. Here, the blog is a general term for sites such as personal diaries on the web, and is also called a web blog. Since FIG. 3 is a web search screen, there is no underline in the character portion of the web.

  The registration screen 27 is a screen for the registrant 6 who created the content to register his / her website on the search site 2. As shown in FIG. 4, the name and contact information of the registrant 6 of the content to be registered, etc. There is an area for entering information related to general content posted on the site, such as information related to registrants 6 and the URL of the website to be registered, whether or not to register with a robot-type search engine, hope for directory registration, and adult ratings. It has been.

  For each content, the name of the content, the type of content, the URL where the content is registered, and the comment of the registrant 6 on the content are registered. With regard to content registration, the entry field can be scrolled by the scroll bar 35, and many contents can be registered. A registration button 36 is provided above the registration screen 27. The search server 1 receives registration from the registrant 6 and many other registrants in advance, and stores the metadata 22 and the content evaluation value 23 in the content DB 21. The search result screen 28 will be described later.

  Next, the operation of the search server 1 of the first embodiment will be described with reference to FIGS. Here, a case where “panda” is used as the search term 24 will be described. The user 4 causes the user terminal 5 to display the top page 26 of the search site 2 and inputs “Panda” as the search word 24 in the input area 29. When the search button 30 on the top page 26 is clicked, a site search request is generated for the search server 1 (S1).

  As described above, when the search word 24 is input to the input area 29 and the search button 30 is clicked, the check program 20 first checks whether the site search request is valid (S2). This legitimacy check is called sanitization, and is intended to eliminate unauthorized access such as access aimed at attacking the server. For example, when there is an unauthorized access such as so-called cross-site scripting, the check program 20 performs error processing such as denying a use request (S11), assuming that the access is an unauthorized access (NO in S2).

  Next, a search in the content DB 21 is performed by the extraction program 14 (S3). Specifically, for each site registered in the content DB 21, it is searched whether there is a corresponding content including the character string “panda” as the search word 24 in the metadata 22. Then, after searching all the contents in the site (Yes in S4), it is checked whether or not the corresponding contents 25 exist in the site (S5).

  If the corresponding content 25 exists (Yes in S5), the evaluation calculation value calculation program 15 calculates the evaluation calculation value for each site (S6). In the first embodiment, an evaluation calculation value is obtained in a form that considers the number of corresponding contents of the site on the basis of an average value of the evaluation values 23 of the contents. Specifically, the average value of the evaluation values 23 of the corresponding contents 25 is calculated, the number of contents 25 is counted for each site, and the number of corresponding contents 25 is counted as a predetermined number (in this embodiment, The evaluation calculation value is obtained by multiplying 0.5) and adding to this average value.

  In this embodiment, since there are five panda photos on the home page 9 of the registrant 6, the number of corresponding contents is five. For example, if the content evaluation values are 90 points, 80 points, 80 points, 40 points, and 30 points, the average value of these five evaluation values is 64 points. The result is 2.5, which is obtained by multiplying the number of corresponding contents 5 by 0.5 and becomes 66.5 points. At the same time, the total evaluation calculation value calculation program 16 counts the total number of contents in each site (9 in this embodiment), and calculates the total evaluation calculation value using the same calculation formula as the evaluation calculation value calculation program 15 (S7). ).

  On the other hand, if the corresponding content 25 does not exist in the searched site, it is checked whether or not the search of all the sites in the content DB 21 is completed. If there is a site that has not been searched yet (No in S8), the metadata 22 in the content DB 21 is searched (S3 to S7).

  When the search of all sites in the content DB 21 is completed (Yes in S8), the sort program 17 sorts the sites based on the evaluation calculation values (S9). Specifically, the first condition for sorting is set as an evaluation calculation value, the second condition is set as the number of corresponding contents, the third condition is set as a total evaluation calculation value, and the fourth condition is set as the total number of contents of the site. Then, the list of sites sorted on the user terminal 5 is displayed on the search result screen 28 by the display program 18 (S10).

  As shown in FIG. 6, the search result screen 28 is a screen on which a result of a search performed by the search word 24 input in the input area 29 of the top page 26 is displayed. FIG. 6 shows a search result when the search word 24 is “panda”. In the search result screen 28, the site name 37 of the site where the content including the characters “panda” is stored in the metadata is displayed. In addition, four numbers are displayed on the right side of the site name 37. These numbers are the corresponding content number 38, the evaluation calculation value 39 of the corresponding content, the total content 40 in the site, and the total evaluation calculation value 41 of all the content in the site from the left side. A simple description 42 of the site is displayed below the site name 37. This descriptive text 42 is the content entered by the registrant 6 as a comment on the registration screen 27.

  In FIG. 6, the website ranked “Panda Site” is ranked first, the corresponding content number 38 is “100”, the evaluation calculation value 39 is “93”, and the total content number 40 is “ 300 ”and the total evaluation calculation value 41 is“ 82 ”. This site has a large number of corresponding contents 38 and the evaluation calculation value 39 is very high. Therefore, the user 4 can first browse the “Panda site”, which is a site including a large number of highly evaluated panda-related contents.

  Also, the website ranked “Panda World” is ranked second, and the number of corresponding contents 38 is “10”, but the evaluation calculation value 39 is “85”. The total number of contents 40 is “20”, and the total evaluation calculation value 41 is “82”. This site has a small number of content, but the evaluation is high. Therefore, the user 4 can recognize the site as a site having highly evaluated content regarding the panda.

  In the following, the number of corresponding content 38 is “20” in the third site, and there are more corresponding content 38 than the second site, but the evaluation calculation value 39 is “72” from the second site. Because it is low, it is ranked third. The fourth-ranked site has a corresponding content number of 38, which is as small as “4”. However, since the evaluation calculation value is “70”, the rank is higher than that of the fifth-ranked site. The fifth-ranked site is the homepage 9 registered by the registrant 6 in the web server 8, the corresponding content 38 is “5”, and the evaluation calculation value 39 is rounded to “67”. The total number of contents 40 is “9”, and the total evaluation calculation value 41 of the contents is “61”.

  As described above, in the search server 1 according to the first embodiment, the number of corresponding contents 25 in the searched site on the search result screen 28, and the evaluation calculation value of the corresponding contents 25, that is, the entire corresponding contents 25. You can know what kind of evaluation is done. It is also possible to know the total number of contents in the site and the total evaluation calculation value thereof. Thereby, it is possible to know what is the scale of the site having the corresponding content 25 and what is the evaluation of all the contents existing on the site. For this reason, since the user 4 can browse from a site having highly evaluated content, there is a high possibility that the user 4 can quickly find the content or site desired by the user 4.

  Then, when the user 4 looks at each site displayed on the search result screen 28 in order from the top and clicks on the “zoo photo” site ranked fifth, the home page 9 of the registrant 6 is displayed. Since it is displayed on the user terminal 5, the user 4 can browse the content registered by the registrant 6.

  Next, a modification of the first embodiment will be described with reference to FIG. The first modification is characterized in that the sort program 17 does not rearrange the sites based on the evaluation calculation values, but rearranges the sites based on the number of corresponding contents 25. Then, on the search result screen 28 ′, “Panda World”, which was second in the first embodiment, is ranked third, and “China Panda”, which was third, is second. The 4th place “Panda House” is ranked 5th, and the 5th place “zoo photo” is ranked 4th. Since the first embodiment and the first modification differ depending on whether the user 4 places importance on the evaluation calculation value or the number of corresponding contents, the user 4 may be able to select.

  Next, a second embodiment of the present invention will be described with reference to FIGS. In the second embodiment, before the site search is performed by the search term by the user, the content DB is searched with the predicted search term predicted in advance, and the search result is stored in advance, When a search is actually performed, the result of the search performed in advance is shown to the user.

  As shown in FIG. 8, in the search server 1 ′ in the second embodiment, a second extraction program 51 is stored in the hard disk 12 ′, and a predicted search word table 52 and search results are stored in the content DB 21 ′. 53 is stored. The predicted search word stored in the predicted search word table 52 is created in anticipation of a search word that will be searched by the user 4. The extraction program 14 ′ searches the metadata 22 in the content DB 21 ′ with the predicted search word stored in the predicted search word table 52. The sort program 17 ′ sorts the sites based on the evaluation calculation value and the total evaluation calculation value, and stores the sorting result, the evaluation calculation value, and the total evaluation calculation value in the search result 53.

  The predicted search word stored in the predicted search word table 52 is specifically created as follows. First of all, from the metadata registered on the registration screen 27 shown in FIG. 4, for example, the name of the registrant 6, the name of the content, the comment of the registrant 6 on the content, etc., using a dictionary or morphological analysis. Extract words and use them as expected search terms. This morpheme analysis is a technique for extracting a word or the like by decomposing metadata into morphemes using a dictionary (word list).

  The second extraction program 51 determines whether a predicted search word corresponding to the actual search word 24 exists in the predicted search word table 52 when a search is performed by the search word 24 from the user 4 via the Internet 3. It is a program that searches whether or not, and extracts the predicted search word if it exists. Further, in the second embodiment, when the actual search word 24 does not exist in the predicted search word table 52, the extraction program 14 ′ is applicable using the actual search word 24 as in the first embodiment. The content 25 is extracted. Since other configurations are the same as those of the first embodiment, the same configurations as those of the first embodiment are denoted by the same reference numerals, and detailed description thereof is omitted.

  Next, the operation of the search server 1 ′ of the second embodiment will be described with reference to FIGS. 9 and 10. First, a search performed in advance will be described with reference to FIG. In the search server 1 ′ according to the second embodiment, a predicted search word expected to be searched by the user is stored in the predicted search word table 52, and the search process is performed periodically (for example, once a day). . Specifically, when the search processing start time is reached (Yes in S21), the extraction program 14 'searches for the metadata 22 in the content DB 21' using the predicted search word (S22).

  Then, as in the first embodiment, whether or not there is the corresponding content 25 for each site is searched (S23). If the corresponding content 25 exists (Yes in S24), the evaluation calculation value calculation program 15 applies it. The content evaluation calculation value is calculated (S25), and the total evaluation calculation value calculation program 16 calculates the total evaluation calculation value for all the contents of the site (S26). After all sites have been searched (Yes in S27), the sort program 17 'sorts the sites based on the evaluation calculation values (S28). Then, after the search by all the predicted search words is performed (Yes in S29), the sorting result, the evaluation calculation value, and the total evaluation calculation value are stored in the search result 53 (S30).

  Next, the search actually performed by the user 4 will be described with reference to FIG. When the user 4 actually performs a search using a search word (S31), first, the validity of access is checked (Yes in S32), and then the second extraction program 51 determines the actual search word and the expected search word. Contrast with. If the actual search word matches the expected search word (Yes in S33), the search, calculation, and sorted results are extracted from the search result 53 (S34), and the display program is displayed. 18 is displayed on the user terminal 5 (S35).

  On the other hand, if the actual search word and the expected search word do not match (No in S33), the extraction program 14 ′ searches the metadata 22 in the content DB 21 ′, and the evaluation value calculation program 15, Then, the evaluation value and the total evaluation value are calculated by the total evaluation value calculation program 16, and an actual search process is performed (S36). Then, the display program 18 displays the result on the user terminal 5 (S35).

  As described above, in the second embodiment, the search based on the predicted search word, the extraction of the corresponding content 25, and the rearrangement of the sites are performed in advance, so that when the user 4 actually performs the search, the search is quickly performed. The result can be displayed on the user terminal 5.

  In the second embodiment, as shown in FIG. 11, the search result screen 54 displayed by the display program 18 displays an evaluation calculation value 55 and a total evaluation calculation value 56 as an image in a bar graph in two upper and lower stages. ing. The upper graph shows the score of the evaluation calculation value 55, and the file type is displayed on the graph. For example, in the site “Panda site” ranked first, it can be seen that the file type is “JPG”, that is, an image file, and “MOV”, that is, a moving image file. The area of the file type is divided on the graph, and the ratio of the number of contents of each file type is displayed. In the first site, it can be seen that image files account for about 70% and movie files account for about 30%.

  The lower graph represents the score of the total evaluation calculation value 56, and the number 57 of the corresponding content 25 and the total number 58 of the contents in the site are displayed on the graph. In the first site, the number of corresponding contents is 100, and the total number of contents is 300. Therefore, in this graph, it is possible to recognize at a glance how much of the total content the corresponding content 25, that is, the corresponding content 25 having the search word as metadata.

  The meaning of the graph and the meaning of the numbers displayed on the graph can be determined by the user 4 by clicking the link 59 “* Explanation of graph” on the screen of FIG. Can be understood by looking at the screen (not shown) described.

  Next, a modification of the second embodiment will be described with reference to FIG. In this modified example, the search result screen 54 ′ displays the name of the site, a brief description thereof, the number and type of the corresponding content 25, and the evaluation for each content. The length of the bar graph denoted by reference numeral 60 in FIG. 12 represents the number of corresponding contents 25. For example, “Panda site” ranked first has 100 corresponding contents, and “China Panda” ranked second has 20 corresponding contents. The length of the bar graph is 1/5 that of the first bar graph.

  Further, the shading of the bar graph in FIG. 12 indicates the difference in color, and indicates the evaluation value of each content. The color classification is such that the evaluation of the dark portion on the left side is high, and the evaluation decreases toward the right side. Has been. Specifically, the darkest part at the left end is red and the evaluation value is 90 or more and 100 or less, the next darkest part is orange and the evaluation value is 80 or more and less than 90, and the next darkest part is yellow and the evaluation value Are 70 or more and less than 80, and are yellowish green, green, blue-green, patina, blue, deep purple, and black as the evaluation becomes lower.

  Thus, the user 4 can intuitively select a site with a high evaluation by expressing the number of the corresponding contents 25 by the length of the graph and expressing the evaluation value of the corresponding content 25 by the color. Therefore, since the user 4 can browse from highly evaluated sites and highly evaluated contents, there is a high possibility that the user's desired contents can be quickly reached.

  Next, a second modification of the second embodiment will be described with reference to FIG. In the second modified example, the search result screen 54 ″ has a total site evaluation indicating the total evaluation calculation value calculated by the total evaluation calculation value calculation program 16 on the vertical axis, and the total number of contents included in the site is displayed. A bubble chart is displayed on the horizontal axis 61. The size of the circle 61 in Fig. 13 indicates the number of the corresponding contents 25, and the number of the corresponding contents 25 is a number in the lower right part of the circle 61. Is represented (reference numeral 62).

  In the second modification, when the user 4 moves the mouse cursor 63 in the circle 61 on the user terminal 5, the name and brief description of the site, the number of the corresponding contents 25, the number of the corresponding contents 25 The evaluation, the total content number of the site, and the total evaluation calculation value of the content are displayed on the display window 64. Further, the color of each circle 61 is color-coded according to the evaluation calculation value of the corresponding content 25. When the user 4 clicks a circle on a predetermined site, the site is displayed on the user terminal 5.

  In this way, by expressing the number of corresponding contents 25 by the size of the circle 61 and the evaluation calculation value of the corresponding contents 25 by color, the user 4 can visually compare between the sites. You can select a highly rated site. Further, when a site is represented by a bubble chart as shown in FIG. 13, many sites can be displayed on one screen, so that the number of times the user 4 changes the screen can be reduced.

  In the first and second embodiments, the search servers 1 and 1 'are described as examples. However, the present invention is not limited to this, and the ranking server disclosed in the Japanese Patent Application No. 2004-37363 application by the present inventor. The function of the search server of the present invention may be added. In the search in the application, the content itself is searched and the search results are displayed in the order of evaluation. Therefore, it is also possible to sort the sites by the corresponding contents by the function of the search server of the present invention and display not only the contents list but also the results of the site search.

  In the evaluation calculation value calculation program 15, various statistical methods may be used as a method for calculating the evaluation calculation value. Specifically, in addition to the above calculation method, an average value of content evaluation values may be simply used as an evaluation calculation value, and a positive value of a variance value (a value obtained by subtracting the average value from each evaluation value) is added. A value obtained by multiplying a predetermined count may be added to the average value, or a value obtained by multiplying the standard deviation by a predetermined count may be added to the average value.

  In addition, in the above-described embodiment, the registrant 6 registers in advance the URL of the site where the content stored in the content DBs 21 and 21 ′ exists. However, the URL is not limited to this and is estimated from the content URL. May be. For example, if the URL is “http://www.example.com/~username/images/photo.jpg”, it is “http://www.example.com/” or “http: / /www.example.com/~username/ ", the registrant 6 may be asked to present and check these expected URLs, saving the input effort.

  Further, in the extraction programs 14 and 14 ′, when the number of contents to be extracted is small, contents having words related to the search word 24 or the predicted search word may be extracted. For example, if there is little content having the word “cloud”, content having the word “sky” related to “cloud” may also be extracted. In this case, since the word “blue sky” is expected to be less related to “cloud”, the word “blue sky” is excluded.

  In the second embodiment, the second extraction program 51 compares the actual search word and the predicted search word. However, the approximation is not selected without making this comparison strict. The search result of the predicted search word may be presented to the user 4. For example, even if content with metadata “bike” exists, it cannot be searched with the search term 24 “motorcycle”, but “bike” is a synonym for the word “motorcycle”. Display search results based on expected search terms.

  Further, the search term 24 input by the user 4 to the top page 26 of the search site 2 shown in FIG. 3 is stored, the statistics of the search term 24 are taken, and the metadata of the content is retrieved while being searched a predetermined number of times or more. A search term 24 that is not included may be extracted, and such a search term 24 may be used as a predicted search term. In this case, it corresponds to a synonym, a synonym, a foreign language, or an abbreviation of the predicted search word, and is associated with the word actually included in the metadata and is extracted in this way when the extraction program 14 'performs extraction. The search result 53 based on the word associated with is stored in the content DB 21 ′.

  By comparing the search terms that are somewhat ambiguous in this way, a site having metadata of search terms close to the search terms specified by the user 4 can be quickly displayed on the user 4. In the second embodiment, the predicted search word table 52 and the search result 53 are stored in the content DB 21 '. However, the present invention is not limited to this, and it may be stored in the search server 1'.

  In the new registration screen in FIG. 4, the adult rating is registered for each site. However, the present invention is not limited to this, and an adult rating may be registered for each content. In the embodiment shown in FIG. 11, the evaluation value of each corresponding content 25 is not displayed. However, as shown in FIG. 12, the evaluation value of each corresponding content 25 may be represented by a color.

  In the second modification of the second embodiment, search results are displayed two-dimensionally, but may be displayed three-dimensionally. For example, an evaluation calculation value may be taken for the depth, and a site with a higher evaluation calculation value may be displayed on the front side, or a site that has been updated by taking the update date and time of the site in the depth will be displayed on the front side. It may be. Further, in FIG. 13, the number of corresponding contents 25 is represented by the size of a circle. However, the shape is not limited to this, and the shape is changed depending on the update date and time, such as a polygon such as a rectangle or a pentagon, a star, or the like. The number of corresponding contents 25 may be represented.

Explanatory drawing which shows the example of a connection with the internet of the search server which is an example of embodiment of this invention. The block diagram which shows the functional structure of the search server of 1st Embodiment. Explanatory drawing which shows the example of the top page of the search site in a search server. Explanatory drawing which shows the contents of a content registrant's site and content registration. The flowchart which shows the action | operation of the search server of 1st Embodiment. Explanatory drawing which shows the search result in the search server of 1st Embodiment. Explanatory drawing which shows the search result of a 1st modification. The block diagram which shows the functional structure of the search server of 2nd Embodiment. The flowchart which shows the action | operation of the search server of 2nd Embodiment. The flowchart which shows the action | operation of the search server of 2nd Embodiment. Explanatory drawing which shows the search result in the search server of 2nd Embodiment. Explanatory drawing which shows the search result of the modification in 2nd Embodiment. Explanatory drawing which shows the search result of the 2nd modification in 2nd Embodiment.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Search server, 3 ... Internet (network), 4 ... User, 9 ... Homepage (web site), 14 ... Extraction program (extraction means), 15 ... Evaluation calculation value calculation program (evaluation calculation value calculation means), 17 ... sort program (sort means), 18 ... display program (display means), 21 ... content DB, 22 ... metadata, 23 ... content evaluation value, 25 ... corresponding content.

Claims (9)

For a plurality of contents existing in a plurality of sites, a content database storing metadata having information on the contents including at least a content evaluation value and a URI of the site where the contents exist;
An extraction means for extracting the content including the search word in the metadata from the content database when a search is performed by a user via a network;
An evaluation calculation value calculating means for calculating evaluation calculation values of all corresponding contents for each site by a predetermined calculation formula based on the evaluation value of the corresponding content;
Sorting means for sorting each site based on the evaluation calculation value or the number of the corresponding contents;
A search server, comprising: a display unit configured to display a sorted site together with the evaluation calculation value or the number of the corresponding contents so that a user can view the sorted site.   The evaluation calculation value calculation means calculates a total evaluation calculation value by a second calculation formula based on the evaluation of all contents at each site, and the display means calculates the total evaluation calculation value and the total number of contents at the site. The search server according to claim 1, further displayed. For a plurality of contents existing in a plurality of sites, a content database storing metadata having information on the contents including at least a content evaluation value and a URI of the site where the contents exist;
First extraction means for extracting the corresponding content including the predicted search word in the metadata from the content database using a predicted search word expected to be searched by a user;
An evaluation calculation value calculating means for calculating evaluation calculation values of all corresponding contents for each site by a predetermined calculation formula based on the evaluation value of the corresponding content;
Sorting means for sorting each site based on the evaluation calculation value or the number of the corresponding contents;
A result storage means for storing the sorted result, the evaluation calculation value, and the number of the corresponding contents;
A second extraction means for extracting the predicted search word corresponding to the actual search word from the result storage means when a search by an actual search word is performed via a network from a user;
A display means for displaying a sorted result site corresponding to the predicted search word extracted by the second extraction means so that the user can view it together with the evaluation calculation value or the number of the corresponding contents. Search server.   When there is no predicted search term corresponding to the actual search term, the first extraction unit extracts the content including the actual search term in the metadata from the content database, and the evaluation calculation value calculation unit Calculates an evaluation calculation value of all corresponding contents for each site by a predetermined calculation formula based on the evaluation value of the corresponding content, and the sorting means uses the evaluation calculation value or the number of corresponding contents as a reference for each site. 4. The search server according to claim 3, wherein the site is sorted so that a user can view the sorted site together with the evaluation calculation value or the number of the corresponding contents. 5.   The evaluation calculation value calculation means calculates a total evaluation calculation value by a second calculation formula based on the evaluation of all contents at each site, and the result storage means calculates the total evaluation calculation value and the total number of contents at each site. 4. The search server according to claim 3, wherein the display means further displays the total evaluation calculation value and the total number of contents in the site.   The search server according to claim 1, wherein the display unit displays the evaluation calculation value and the number of the corresponding contents as an image.   The search server according to claim 2, wherein the display unit displays the total evaluation calculation value and the total number of contents in each site as an image.   The said display means displays the said evaluation calculation value, the said total evaluation calculation value, the evaluation value of a content, the number of applicable content, or the total number of content by a color, The any one of Claim 1 thru | or 7 characterized by the above-mentioned. Search server described in.   The search server according to claim 1, wherein the display unit further displays the type of content.

Images