JP2010225051A - Content meta information impartment device and method thereof, and content retrieval device and method thereof - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a content meta information impartment device allowing utilization of a situation that present retrieval is about to be done for sharing of present and future meta information. <P>SOLUTION: The content meta information impartment device includes: a content meta information holding part holding content meta information; a browsed history information holding part holding browsed history information of a content; and a dynamic content meta information impartment part reading a past history corresponding to a browsed situation from the browsed history information holding part when the prescribed past browsed situation of the content is imparted, calculating similarity between the contents in the browsed situation based on the read browsed history information, imparting the content meta information held in the content meta information holding part as the content meta information of the opposite party between the contents decided that they are similar to each other based on the similarity, or performing processing of increasing weight of the content meta information corresponding to each content according to the similarity. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a content meta information providing apparatus, a method and a program thereof, and a content search apparatus, a method and a program thereof, which share meta information between contents according to a search user's information and a search context. About.

  There are various methods for retrieving necessary information from a large amount of information stored in a computer or the like. In general, when performing a search, a keyword that is considered to be closely related to necessary information is given to the system as an input. On the other hand, in order to extract content based on the received keyword, the search system holds a content index. The content index is obtained by performing analysis such as morphological analysis on the content, and is retained as a term. The content index can be checked for a match with the input keyword. When executing the search, contents whose terms in the index match the input keyword are extracted.

  The term held in the content index may be given by analyzing the content. However, there is a case where meta information indicating the content content provided by the content creator is used as a term. Since the meaning of the meta information is given by the creator of the content, the meta information has a high content of meaning indicating the contents of the content, and by using this meta information for the search, a more accurate search is performed. (See Patent Document 1).

  The use of meta information is premised on the existence of meta information corresponding to content. However, it is generally expensive to manually add meta information to content, and there are many cases where there is no meta information corresponding to the content.

  As a technique to solve the above problem by sharing meta information between content that has corresponding meta information and content that does not have corresponding meta information, collaborate using history information that has been browsed. A technique for performing filtering has been proposed (see Patent Document 2).

JP 2002-334257 A JP 2003-316824 A

  However, in the collaborative filtering method used in Patent Document 2, all the content browsing history is used, so that the current search is being performed in sharing the current and future meta information. The search result based on the meta information does not reflect the search condition and the state at the time of the search.

  Therefore, the present invention utilizes the situation where the current search is about to be performed for sharing the current and future meta information, and the search result based on the meta information reflects the search condition and the state at the time of the search. It is an object of the present invention to provide a content meta information assigning device, a method and a program thereof, and a content search device, a method and a program thereof.

  According to the present invention, a content meta information holding unit that holds content meta information (hereinafter referred to as “content meta information”), a browse history information holding unit that holds content browse history information, When a predetermined past browsing situation is given, a past history corresponding to the browsing situation is read from the browsing history information holding unit, and based on the read browsing history information read, The degree of similarity between contents is calculated, and the contents meta information held in the contents meta information holding unit among the contents determined to be similar to each other based on the degree of similarity A dynamic code that performs processing to increase the weight of the content meta information corresponding to each content according to the similarity or given as content meta information. Content meta information providing device, characterized in that it comprises a Tentsumeta information adding unit, is provided.

  Further, according to the present invention, a content meta information holding unit that holds content meta information (hereinafter referred to as “content meta information”), a browse history information holding unit that holds content browse history information, In the content meta-information providing method in the content meta-information providing apparatus comprising: when a predetermined past browsing situation of the content is given, a past history corresponding to the browsing situation is read from the browsing history information holding unit Based on the read history information that has been read and read, the similarity between the contents in the browsed state is calculated, and the content determined to be similar to each other based on the similarity Each content meta information held in the meta information holding unit is assigned as the content meta information of each partner, or similar Content meta information attaching method characterized by comprising a dynamic content meta information adding step of increasing processing weight for the content meta information corresponding to each content according to is provided.

  Furthermore, according to the present invention, a content meta information holding unit that holds content meta information (hereinafter referred to as “content meta information”), a browse history information holding unit that holds content browse history information, When a predetermined past browsing situation of content is given, a past history corresponding to the browsing situation is read from the browsing history information holding unit, and the browsing history information is read based on the read browsing history information read. The similarity between the contents in the situation is calculated, and the content meta information held in the content meta information holding unit is determined between the contents determined to be similar to each other based on the similarity. A process of giving the content meta information of the other party or increasing the weight of the content meta information corresponding to each content according to the similarity is performed. Program for causing a computer is provided as the content meta information providing device, characterized in that it comprises a specific content meta information adding unit.

  According to the present invention, the current search is performed for sharing current and future meta information, and the search result based on the meta information reflects the search condition and the state at the time of the search. It becomes possible.

It is a block diagram which shows the structure of the content search device by embodiment of this invention. It is a figure which shows an example of the data preserve | saved at the content meta information holding | maintenance part shown in FIG. It is a figure which shows an example of the data preserve | saved in the browsing history information holding | maintenance part shown in FIG. It is a block diagram which shows the structure of the content search apparatus by Example 1 of this invention. It is a sequence diagram which shows operation | movement of the content search apparatus by Example 1 of this invention. It is a block diagram which shows the structure of the content search apparatus by Example 2 of this invention. It is a figure which shows an example of the data currently hold | maintained at the temporary content meta-information holding part shown in FIG. It is a sequence diagram which shows operation | movement of the content search apparatus by Example 2 of this invention. It is a figure which shows an example of the similar content information in Example 1 and 2 of this invention.

  The best mode for carrying out the present invention will be described below in detail with reference to the drawings.

  A block diagram of the system of the present invention is shown in FIG. The system of the present invention includes a content meta information holding unit 101, an inter-content similarity holding unit 103, a browsed history information holding unit 105, a search interface unit 107, a search unit 109, a dynamic content meta information adding unit 111, a static content It includes an inter-similarity calculation unit 113, a static content meta information adding unit 115, and a control unit 117.

  The search interface unit 107 inputs search conditions (search keyword, logical expression for combining search keywords, narrowing conditions (content extension, creation date, creator, category, tag, etc.)) for searching for content. , An input unit 107-1 for passing the input search condition to the control unit 117, and an output unit 107-2 for displaying the search result received from the control unit 117.

  The system shown in FIG. 1 can be realized by hardware, but can also be realized by the computer reading and executing a program for causing the computer to function as the system shown in FIG.

  The content meta information holding unit 101 is used as a search index when searching for content, and the weight of the read meta information that matches the keyword is used for ordering (sorting) the content.

  The browsing history information holding unit 105 holds history information obtained by browsing the content provided by the system by a user of the system. In other words, the browsed history information holding unit 105 searches for each content, the search condition when the content hits the search, and the significance of browsing (whether browsing is performed) for each searcher meta information (title, department, etc.). Keep the percentage of.

  The static content similarity calculation unit 113 obtains the similarity between the contents based on the information held by the browse history holding unit 105.

  The inter-content similarity holding unit 103 holds the inter-content similarity calculated by the static content similarity calculating unit 113.

  The static content meta information adding unit 115 performs processing for sharing meta information between similar contents based on the information held by the inter-content similarity holding unit 103 and the information held by the content meta information holding unit 101. Do.

  The static content meta information adding unit 115 is assumed to be used when updating the meta information of the content meta information holding unit 101 offline (processing in a batch not at the time of search execution). When information is already stored in the browsing history information holding unit 105, but the data is not yet held in the content similarity holding unit 103 and the content meta information holding unit 101, the static content meta information adding unit 115 and the static content similarity calculation unit 113 are used to update the meta information held in the content meta information holding unit 101. The static content meta information adding unit 115 and the static content similarity calculating unit 113 are activated by the system administrator when starting the system of this embodiment. Whereas the dynamic content meta-information adding unit 111 uses a history in a situation that matches the search meta-information (searcher meta-information, context information (such as time and place at the time of search)), static content meta-information The granting unit 115 updates the data held in the content meta information holding unit 101 using all the histories. The static content similarity calculation unit 113 is called from the static content meta information adding unit 115 at that time, but is also called from the dynamic content meta information adding unit 111. The static content meta information adding unit 115 is not used once the system is in operation.

  The dynamic content meta-information adding unit 111 extracts, from the browsed history information holding unit 105, contents that are close to each other in the past history part that matches the search condition and search meta information when using the search. Further, the dynamic content meta information adding unit 111 obtains the inter-content similarity that further enhances the similarity between the extracted contents in addition to the similarity between the contents held in the inter-content similarity holding unit 103. The value of the meta information in the content meta information holding unit 101 is updated according to the degree of enhancement between the contents whose inter-content similarity is enhanced.

  The search unit 109 acquires corresponding content from the keywords included in the search conditions acquired by the search interface unit 107 and rearranges them in an appropriate order.

  As shown in FIG. 2, the content meta information holding unit 101 according to the present embodiment holds a meta information group indicating the ID of the content such as a URL and a property such as the attribute of the content for each content. Yes.

  In the browse history information holding unit 105 of this embodiment, as shown in FIG. 3, a matrix formed by search meta information and content is prepared for each meta item related to search. In the matrix, a value indicating whether each document is necessary for the searcher in the condition of each meta item at the time of search is entered. In this example, the matrix is represented by a binary value indicating whether the document is viewed or not. Thus, in this example, the values in this matrix are values that indicate whether each document was needed by the searcher.

The values in the matrix of FIG. 3 are the significance of browsing, but when the number of browsing is regarded as the significance, the values in the table of FIG. These values are normalized between 0 and 1. By normalizing, there is a merit that the influence of the difference in the accumulation state of the history in the item of the search meta information can be minimized when the similarity is obtained. When normalization is performed, in the browsed history information holding unit 105, for each item of the search meta information, the maximum value and the minimum value of the significance (in this case, the number of browsing) of all contents in the item are set to 1 and 0, respectively. Adjust so that In particular,
Significance after normalization = (Significance before normalization-Minimum value of all content in the item) / (Maximum value of all content in the item-Minimum value of all content in the item))
A method using an expression such as If normalization is performed, normalization is performed in advance in batches at the timing when the browsing history information holding unit 105 is updated, or normalized for each item of search meta information scanned at the time of search execution. do.

  In addition to the number of hits in the search results so far, you can also consider the percentage of actual accesses, but in that case, in the matrix, in addition to the number of views, the number of hits in the search Two values are required. In this case, it is assumed that the history information of the number of times the content search has been hit and the number of times the hit content has been browsed is given from the outside.

  The similarity calculation part 113 between static contents of this embodiment acquires the set of content and its significant value for every search meta information from the browsing history information holding | maintenance part 105, and uses that information, and most is all. The similarity between contents is calculated. As a method of calculating the similarity, an existing technique for obtaining a similarity between vectors (for example, an inner product of vectors) using a significant value for each content search meta information in the browse history information holding unit 105 as a feature vector of the content is used. . The calculated similarity between contents is held in the content similarity holding unit 103.

  An inter-content similarity holding unit 103 according to the present embodiment holds an inter-content similarity calculated by the static inter-content similarity calculating unit 113 and has a function of passing the information as necessary. It is.

  The static content meta information adding unit 115 and the dynamic content meta information adding unit 111 according to the present embodiment read the similarity between the contents from the inter-content similarity holding unit 103, and the similarity between the contents is below a certain threshold. If it is larger, the content meta information holding unit 101 is made to share the meta information held by the contents according to the similarity.

  The dynamic content meta information adding unit 111 according to the present embodiment is a diagram regarding search meta information that matches the key for each content from the browsed history information holding unit 105 using the search meta information acquired from the user at the time of search as a key. The value shown in FIG. 3 (relationship between the name of the search meta information and the content (representing the rate at which the content is browsed when the search meta information appears, etc. Acquire information indicating association)). Based on the acquired value, similar content sets (two contents) are selected when search meta-information as a key is used. For example, there is a content that is viewed similarly when searched by the keyword “mobile phone” even if the browsing history is not very similar between certain contents. Here, the similarity tendency of the browsing history information under such a specific search condition is examined. For content sets determined to be similar, a value obtained by enhancing the content similarity between contents acquired from the inter-content similarity holding unit 103 is acquired. The content meta information holding unit 101 is referred to the content set with the enhanced similarity, and the content meta information shared between the contents is enhanced according to the enhanced similarity.

  The similarity between contents is represented by one numerical value. When calculating the similarity, a numerical value (that is, a plurality of values) for each content meta information as a key is used. As an example, the item of content meta information as a key is the search keyword = SI, searcher title = section manager, searcher affiliation department = sales department in FIG. 3, and the difference threshold is the average value of each search meta information item. When the average of the distance from the threshold is used, the threshold value of the term “SI” is (| (2/3) −1 | + | (2/3) −0 | + | (2/3) −1 |) / 3 = 4/9, similarly, the threshold of the section “section manager” is 4/9, and the threshold of the section “sales department” is 0. The similarity between the document A and the document B is (number difference between the meta items is equal to or less than the threshold of the meta item / number of meta items to be compared) = 1/3 = 0.33. The similarity between document A and document C is (3/3) = 1.

  A configuration example of the search system using the minimum necessary modules in the first embodiment will be described.

  FIG. 4 is a configuration block diagram of the first embodiment. The search system 201 according to the first embodiment includes a search interface unit 107, a control unit 117, a search unit 109, a content meta information holding unit 101, a browsed history information holding unit 105, and a dynamic content meta information adding unit 111. Including.

  The search interface unit 107 is the same as that of the embodiment.

  The control unit 117 has a part for acquiring search conditions input by a user of the search system 201 and user meta information such as a user's department and job title, and moves the acquired search conditions and user meta information. To the content meta information adding unit 111. Then, when receiving the notification that the content meta information has been updated from the dynamic content meta information adding unit 111, the control unit 117 sends the search condition input by the user as a search request to the search unit 109. The search result returned from the search unit 109 is passed to the user.

  It is assumed that user meta information is given from an external network. As an apparatus necessary for acquiring user meta information, an “authentication system” for acquiring user personal information is required.

  The search unit 109 includes a search core unit 109-1 and a scoring unit 109-3.

  The search core unit 109-1 accesses the content meta information holding unit 101, acquires content having content meta information that matches the acquired search condition acquired by the control unit 117 in the above operation, and at the same time, for each content, Get the sum of the weights of the meta information group that matches the keyword.

  The scoring unit 109-3 sorts the content acquired by the search core unit 109-1 with reference to the weight of the meta information group, and passes the sorted content to the control unit 117.

  The content meta information holding unit 101 is a storage unit that holds meta information representing content attributes. Each content has a plurality of pieces of meta information (may not have meta information). In addition to meta items, meta information may include a combination of meta items and their values, and a configuration having weights for the combinations. An example of the configuration of the meta information held by the content meta information holding unit 101 is shown in FIG.

  The browsing history information holding unit 105 holds, for each content, a value (valid value) indicating whether the content presented by the search system is significant to the user. The value may be a binary value of 0 or 1 that simply indicates whether the user has viewed the content. In addition, as a value, when a button such as “not helpful” or “not helpful” is installed on the page to be browsed, the ratio of which button is pressed may be adopted. While the above effective values are held for each content, they are also classified for each search condition and for each meta information of a user such as a department or job title.

  An example of the structure of data held in the browse history information holding unit 105 is shown in FIG.

  The dynamic content meta information addition unit 111 includes a dynamic content similarity calculation unit 111-1 and a dynamic content meta information addition core unit 111-3.

  The dynamic content similarity calculation unit 111-1 includes the keyword and user meta information items provided from the dynamic content meta information adding unit 111 in the history held by the browsed history information holding unit 105. The ratio of the number of items in which the difference in significance is equal to or less than a predetermined effective value threshold in the number of significance levels between contents is calculated and returned.

  After receiving the search meta information, the dynamic content similarity calculation unit 111-1 first accesses the browsed history information holding unit 105 and scans all the pairs of contents. At the time of scanning, it is digitized by examining how much the browsing history is similar in the content meta information of the content that matches the received content meta information.

  While the static content similarity calculation unit 113 in FIG. 1 scans all the two contents in advance and calculates the similarity between all the contents, the dynamic content similarity calculation unit 111-1 The similarity between only the search keyword given from the control unit 117 and the content having the search meta information is calculated.

  The dynamic content meta-information adding core unit 111-3 has a portion that refers to the similar content information (see FIG. 9) provided by the dynamic content meta-information adding unit 111. For all items (two content URLs and weight (similarity)) included in the referenced similar content information, the content meta information holding unit 101 refers to the meta information of the content having the content URL in the item. . Relative to the meta information of the content of the referenced pair, the value of the meta information of the meta information that one of the pair has is multiplied by the weight of the meta item and the weight (similarity) referenced in the item. (If there is no value of the meta item to be added to the opposite meta information, the value of the meta item is newly added, and the numerical value calculated by multiplying the weight is set first). By performing this for each of the paired contents, the meta information of these contents is updated. The updated meta information is updated to the content meta information holding unit 101.

  Next, the operation will be described.

  FIG. 5 is a sequence diagram illustrating an example of processing according to the first embodiment.

  First, the control unit 117 that has acquired the search request transmits the acquired search condition and the search meta information, and also transmits a content meta information update request to the dynamic content meta information adding unit 111 (step S301).

  Receiving this, the dynamic content meta information adding unit 111 requests the dynamic content similarity calculation unit 111-1 to calculate the similarity between the contents (step S302). Here, the content for which the similarity is calculated is a combination of the content including the search condition and meta information received by the control unit 117 and all the content.

  Upon receiving the similarity calculation request, the dynamic content similarity calculation unit 111-1 acquires content browsing history information from the browsing history holding unit 105 (steps S 303 and S 304), and calculates similarity information between the content. Then, it returns to the dynamic content meta information adding unit 111 (step S305).

  Receiving it, the dynamic content meta information adding unit 111 passes the similarity between the acquired contents to the dynamic content meta information adding core unit 111-3 (step S306).

  The dynamic content meta-information adding core unit 111-3 that receives the similarity information between the contents rewrites part of the meta information of the content meta-information holding unit 101 (steps S307, S308, and S309).

  After rewriting the meta information, the dynamic content meta information adding unit 111 notifies the calling source control unit 117 that the processing has been completed (step S310).

  The control unit 117 notified of the completion of the process requests the search unit 109 to perform a search together with the search condition (step S311).

  The requested search unit 109 requests the search core unit 109-1 to perform a search (step S312), and the search core unit 109-1 determines the content of the keyword and the term in the index and the value of the meta item. The sum of the weights is extracted from the content meta information holding unit 10 and passed to the search unit 109 (steps S313, S314, and S315). Receiving the extracted content, the search unit 109 passes the received content to the scoring unit 109-3 (step S316).

  The scoring unit 109-3 sorts the search results (such as the title and URL of the extracted content), and returns the search results obtained as a result to the search unit 109 (step S316). The search result is returned to the search system user (step S318).

  The first embodiment aims to improve the recall rate of the search, while the second embodiment aims to improve the precision.

  In the second embodiment, meta information obtained at the time of searching is not made permanent as the entire meta information, but is generated as meta information for searching for the situation at the time of searching.

  In the first embodiment, the content meta information in the content meta information holding unit 101 is updated every time search is executed. In the second embodiment, the content meta information is updated by the content meta information holding unit 101 in the entire content meta information holding unit 101. By temporarily preparing a meta information field (temporary content meta information holding unit 121) that is not reflected in the information and valid only during the search, a history of past search behaviors that are not related to other current situations Eliminate the impact.

  In the second embodiment, content meta information related only to the current situation is shared, thereby aiming to share content meta information that matches the search situation (search condition, use meta information).

  In the first embodiment, since all search behaviors are reflected in the sharing of content meta information, sharing of content meta information has the advantage of advancing quickly, but searching at the time of searching with a plurality of search conditions and a plurality of search meta information When meta information is shared for the same content by action, content meta information under different circumstances (search conditions, search meta information) is mixed, and content meta information is widely shared to reproduce search Although the rate is improved, there is a possibility that the relevance rate is lowered by including content meta information that is not related to the current situation.

  In the second embodiment, therefore, the content meta information that is related only under the search status (search condition, search meta information) is shared, thereby aiming at an improvement in the relevance rate.

  If the meta information of the original content is overwritten as in the first embodiment, when the content hits at the time of search, the search meta data matches the original content meta information and the search condition is matched. Since the given meta information matches the search condition, it is not possible to know whether the search has been hit. Furthermore, the meta information given from the history is given according to the search condition at the time of the search and the similarity in the search meta information, and the content meta information in all searches is added to the content meta information one after another. If this happens, the content meta information may become noise. In the second embodiment, the problem of the first embodiment can be solved by temporarily generating and using a search meta information holding unit (temporary content meta information holding unit 121) separately from the original content meta information holding unit 101 during the search. Resolve.

  FIG. 6 is a configuration block diagram of the second embodiment. The search system 202 according to the second embodiment includes a search interface unit 107, a control unit 117, a search unit 109, a content meta information holding unit 101, a temporary content meta information holding unit 121, a browse history information holding unit 105, A dynamic content meta information adding unit 111 is included.

  When the search unit is requested for content information for holding temporary content meta information, the search unit searches all content information that matches the content URL passed at the time of request for the content held in the content meta information holding unit 101. Get and return.

  The search interface unit 107 is the same as that of the embodiment.

  The search unit 109 includes a search core unit 109-1B and a scoring unit 109-3.

  The search core unit 109-1B accesses the search condition and the search field that matches the search meta information held in the content meta information holding unit 101 and the temporary content meta information holding unit 121, and the search condition acquired from the searcher And content identification information having content meta information including the value of the content meta item that matches the search meta information (including searcher meta information, context information at the time of search, etc.), and at the same time, for each content, the searcher And the sum of the weights of the content meta items that match the search meta information are also acquired.

  The temporary content meta information holding unit 121 is a part capable of holding for each search meta information a function of incorporating content having meta information that matches the search conditions equivalent to the content meta information holding unit 101. A configuration example of data held in the temporary content meta information holding unit 121 is shown in FIG.

  The dynamic content meta-information adding core unit 111-3B has the function of the dynamic content meta-information adding core unit 111-3, but the update destination of the calculated meta information is the content meta-information holding unit. The temporary content meta information holding unit 121 is not 101.

  The meta information at the update destination of the calculated meta information is the content meta information. Further, regarding the association between the search meta information and the content meta information, when it is written, the content meta information is calculated from the search meta information input from the searcher, and input from the searcher on the left side of FIG. The calculated content meta information is written on the right side of the search meta information as it is. When reading at the time of search, the content meta information on the right side corresponding to the search meta information on the left side in FIG. 7 that completely matches the search meta information input by the searcher is read.

  Next, the operation will be described. FIG. 8 is a sequence diagram illustrating processing of the second embodiment.

  First, the control unit 117 that has acquired the search request sends a meta information update request together with the acquired information to the dynamic content meta information adding unit 111 (step S401).

  In response, the dynamic content meta information adding unit 111 requests the dynamic content similarity calculation unit 111-1 to calculate the similarity between the contents (step S402).

  The dynamic content similarity calculation unit 111-1 that has received the similarity calculation request acquires the browsing history information of the content from the browsing history holding unit 105 (steps S 403 and S 404), calculates the similar content information, This is returned to the dynamic content meta information adding unit 111 (step S405). The similar content information is information having zero or more items each of which sets a set of similarities between two URLs determined to be similar. A specific example is as shown in FIG.

  Receiving it, the dynamic content meta information adding unit 111 passes the acquired similarity between the contents to the dynamic content meta information adding core unit 111-3B (step S406).

  The dynamic content meta-information adding core unit 113-1B passes the similar content information received from the dynamic content meta-information adding unit 111 to the search unit together with the search meta information, and the content information for the temporary content meta information holding unit 121 To obtain the content information of the content URL described in the similar content information (all information of the content that is held in the content meta information holding unit 101 and matches the passed content URL). Then, the acquired content information is added to the temporary content meta information holding unit 121 (steps S407, S408, S409).

  After rewriting the temporary content meta information, the dynamic content meta information adding unit 111 notifies the calling source control unit 117 that the processing has been completed (step S410).

  The control unit 117 notified of the completion of the process requests the search core unit 109-1B to perform a search together with the search condition and the search meta information (step S411).

  The search unit 109 that has requested the search passes the search condition and search meta information to the search core unit 109-1B to request a search (step S412), and the search core unit 109-1B searches for the search condition and the search meta information. Is sent to the temporary content meta information holding unit 121 (step S413), and the temporary content meta information holding unit 121 extracts the URL of the content that matches the search condition and the search meta information, and the URL of the content and the value of the meta item Is transferred to the search core unit 109-1B (S414).

  In addition, the search core unit 109-1B sends a search condition to the content meta information holding unit 101 to request a search (step S415), and the content meta information holding unit 101 sets the URL of the content that matches the search condition. The content is extracted and the sum of the weights matching the URL of the content and the value of the meta item is passed to the search core unit 109-1B (S416).

  The search core unit 109-1B passes to the search unit 109 the sum of the weights that match the URL of the content and the value of the meta item passed from the temporary content meta information holding unit 121 and the content meta information holding unit 101 (step S417). .

  The search unit 109 that has received the sum of the weights matching the URL of the content and the value of the meta item passes them to the scoring unit 109-3 (step S418), and the scoring unit 109-3 arranges the search results. Change. The scoring unit 109-3 returns the search result obtained as a result to the search system user via the search unit 109 and the control unit 117.

  The data held in the temporary content meta information holding unit 121 can be deleted.

  The effect of this system is to share the meta information according to the search condition at the time of search, the meta information of the searcher, and the context information in a state where the content to be searched is mixed with and without meta information Thus, it is possible to realize a search according to the situation even for content having no content meta information.

DESCRIPTION OF SYMBOLS 101 Content meta-information holding part 103 Inter-content similarity holding part 105 Browsing history information holding part 107 Search interface part 109 Search part 109-1 Search core part 109-1B Search part 109-3 Scoring part 111 in Example 2 Dynamic content meta information giving unit 111-1 Dynamic content similarity calculating unit 111-3, 111-3B Dynamic content meta information giving core unit 113 Static content similarity holding unit 115 Static content meta information giving Unit 117 Control unit 121 Temporary content meta information holding unit

Claims (35)

A content meta information holding unit for holding content meta information (hereinafter referred to as “content meta information”);
Browsing history information holding unit for holding browsing history information of content;
When a predetermined past browsing situation of content is given, a past history corresponding to the browsing situation is read from the browsing history information holding unit, and the browsing history information is read based on the read browsing history information read. The similarity between the contents in the situation is calculated, and the content meta information held in the content meta information holding unit is determined between the contents determined to be similar to each other based on the similarity. A dynamic content meta information adding unit that performs processing to increase the weight of the content meta information corresponding to each content according to the similarity, or as content meta information of the other party;
A content meta information providing apparatus comprising:   2. The content meta according to claim 1, wherein the browse history information of the content indicates a search condition when content for each content hits a search, and a ratio of browsing significance for each searcher meta information. Information grant device.   The browsed history information holding unit is configured to search for each content, each search condition, and each searcher meta information as a significant value by using a ratio when the content is significant for the viewer as a significant value. The content meta information adding device according to claim 1 or 2, wherein the content meta information adding device holds each information.   The browsed history information holding unit holds, for each content, each search condition, and each searcher meta-information, a ratio when the user actually browses among cases presented to the user. The content meta information providing apparatus according to claim 1 or 2, characterized in that the apparatus is provided.   In the dynamic content meta information adding unit, the browse history information holding unit includes the number of both values held for the given item in the search meta information between the two contents, and the value The content meta information according to any one of claims 1 to 4, wherein the content meta information according to any one of claims 1 to 4 is obtained as a similarity by acquiring a number of those having a difference of less than or equal to a threshold in a number of both existing numbers. Granting device.   The content meta according to claim 5, wherein the certain threshold value is an average of significance in each meta item of the browsing history holding unit and a difference in significance of content in the meta item. Information grant device.   The dynamic content meta-information adding unit, when there is some other content whose similarity with the content is equal to or greater than a threshold value for the certain content, assigns the meta-information and weight of the other certain content to the content meta The value obtained by multiplying the similarity and the acquired weight obtained from the information holding unit is added to the weight of the meta information of the certain content, and the meta information possessed by the other certain content is added to the certain content. 7. The content meta information adding apparatus according to claim 1, wherein if there is no content, the value multiplied by the meta information is added to the certain content.   An inter-content similarity calculating unit that calculates the similarity between contents held by the inter-content similarity holding unit, and a content meta that is shared by the content according to the obtained similarity and is held in the content meta-information holding unit The content meta information providing apparatus according to claim 1, further comprising an information adding unit.   The content meta information adding device according to claim 1, wherein when the search condition is given, the content meta information held in the content meta information holding unit is searched using the search condition as a key, and the search A search unit for extracting content meta information whose value matches the condition and meta item value, and when the search condition and search meta information are given, the search condition and the search meta information are passed to the dynamic content meta information adding device. A series of processes for updating the content meta information in the content meta information holding unit, and then acquiring content meta information whose meta condition value matches the search condition from the content meta information holding unit using the search unit A content search apparatus comprising a control unit for performing the above. The dynamic content meta information providing unit further includes:
When a predetermined past browsing situation of content is given, a past history corresponding to the browsing situation is read from the browsing history information holding unit, and the browsing history information is read based on the read browsing history information read. The degree of similarity between the contents in the situation is calculated, and the contents meta information held in the temporary contents meta information holding unit between the contents determined to be similar to each other based on the degree of similarity A process for giving the content meta information of the other party or increasing the weight of the content meta information corresponding to each content according to the similarity,
The search unit excludes the updated content from the search target from the content meta information holding unit, and sets the temporary content meta information holding unit as a search target using the search condition and the search meta information as a key. The content search device according to claim 9.   The search unit includes contents having the same search condition as the given search condition and related meta information for the set of content and meta information in the content meta information holding unit and the temporary content meta information holding unit. The content search apparatus according to claim 10, wherein content having the same is extracted and the content is sorted using the weight of meta information whose meta condition value matches the search condition.   It further comprises an input unit for acquiring a search condition for searching for content and passing the acquired search condition to the control unit, and an output unit for displaying the search result received from the control unit. The content search device according to any one of claims 9 to 11. A content meta information holding unit for holding content meta information (hereinafter referred to as “content meta information”);
Browsing history information holding unit for holding browsing history information of content;
A content meta information providing method in a content meta information providing apparatus comprising:
When a predetermined past browsing situation of content is given, a past history corresponding to the browsing situation is read from the browsing history information holding unit, and the browsing history information is read based on the read browsing history information read. The similarity between the contents in the situation is calculated, and the content meta information held in the content meta information holding unit is determined between the contents determined to be similar to each other based on the similarity. Content meta information comprising a dynamic content meta information adding step for performing processing to increase the weight of the content meta information corresponding to each content according to the similarity, or as content meta information of the other party Grant method.   14. The content meta according to claim 13, wherein the browse history information of the content indicates a search condition when the content for each content hits the search, and a ratio of the significance of browsing for each searcher meta information. Information grant method.   The browsed history information holding unit is configured to search for each content, each search condition, and each searcher meta information as a significant value by using a ratio when the content is significant for the viewer as a significant value. 15. The content meta information adding method according to claim 13, wherein the information is held for each piece of information.   The browsed history information holding unit holds, for each content, each search condition, and each searcher meta-information, a ratio when the user actually browses among cases presented to the user. 15. The content meta information providing method according to claim 13, wherein the content meta information is added.   In the dynamic content meta information giving step, the number of values that are held for the given item in the search meta information between the two contents in the browse history information holding unit, and the value The content meta information according to any one of claims 13 to 16, wherein the content meta information according to any one of claims 13 to 16 is obtained as a similarity by acquiring the number of those having a difference of less than or equal to a threshold in a number of both existing numbers Grant method.   18. The content meta according to claim 17, wherein the certain threshold is an average of significance in each meta item of the browse history holding unit and an average of differences in significance of content in the meta item. Information grant method.   In the dynamic content meta information assigning step, when there is some other content whose similarity to the content is equal to or greater than a threshold, the meta information and weight of the other content are assigned to the content meta information. The value obtained by multiplying the similarity and the acquired weight obtained from the information holding unit is added to the weight of the meta information of the certain content, and the meta information possessed by the other certain content is added to the certain content. 19. The content meta information providing method according to claim 13, wherein if there is no content, the value multiplied by the meta information is added to the certain content. 19.   A content similarity calculation step for calculating a similarity between contents held in the inter-content similarity holding unit, and a content meta for sharing the meta information between the contents according to the obtained similarity and holding the content meta information in the content meta-information holding unit 20. The content meta information providing method according to claim 13, further comprising an information adding step.   14. The content meta information adding method according to claim 13, wherein when the search condition is given, the content meta information held in the content meta information holding unit is searched using the search condition as a key, and the search A search step for extracting content meta information whose value matches the condition and meta item value, and when the search condition and the search meta information are given, the search condition and the search meta information are passed to the dynamic content meta information adding device. A series of processes for updating the content meta information in the content meta information holding unit, and then acquiring content meta information whose meta condition value matches the search condition from the content meta information holding unit using the search unit The content search method characterized by including the control step which performs. In the dynamic content meta information giving step,
When a predetermined past browsing situation of content is given, a past history corresponding to the browsing situation is read from the browsing history information holding unit, and the browsing history information is read based on the read browsing history information read. The degree of similarity between the contents in the situation is calculated, and the contents meta information held in the temporary contents meta information holding unit between the contents determined to be similar to each other based on the degree of similarity A process for giving the content meta information of the other party or increasing the weight of the content meta information corresponding to each content according to the similarity,
In the search step, the updated content is excluded from the search target from the content meta information holding unit, and the temporary content meta information holding unit is set as a search target using the search condition and the search meta information as a key. The content search method according to claim 21, wherein:   In the search step, for the set of content and meta information in the content meta information holding unit and the temporary content meta information holding unit, content having the same search condition as the given search condition and related meta information 23. The content search method according to claim 22, wherein the content having the above is extracted, and the content is sorted using the weight of the meta information whose meta condition value matches the search condition. A content meta information holding unit for holding content meta information (hereinafter referred to as “content meta information”);
Browsing history information holding unit for holding browsing history information of content;
When a predetermined past browsing situation of content is given, a past history corresponding to the browsing situation is read from the browsing history information holding unit, and the browsing history information is read based on the read browsing history information read. The similarity between the contents in the situation is calculated, and the content meta information held in the content meta information holding unit is determined between the contents determined to be similar to each other based on the similarity. A dynamic content meta information adding unit that performs processing to increase the weight of the content meta information corresponding to each content according to the similarity, or as content meta information of the other party;
A program for causing a computer to function as a content meta information providing device. In the content meta-information provision apparatus of Claim 24,
The content browsing history information indicates a search condition when content for each content hits a search, and a percentage of browsing significance for each searcher meta information. A program to make it work.   The browsed history information holding unit is configured to search for each content, each search condition, and each searcher meta information as a significant value by using a ratio when the content is significant for the viewer as a significant value. 26. The program for causing a computer to function as the content meta information adding apparatus according to claim 24 or 25, wherein the program is stored for each piece of information.   The browsed history information holding unit holds, for each content, each search condition, and each searcher meta-information, a ratio when the user actually browses among cases presented to the user. The program for functioning a computer as a content meta-information provision apparatus of Claim 24 or 25 characterized by the above-mentioned.   In the dynamic content meta information adding unit, the browse history information holding unit includes the number of both values held for the given item in the search meta information between the two contents, and the value The content meta information according to any one of claims 24 to 27, wherein the content meta information according to any one of claims 24 to 27 is obtained as a similarity by acquiring the number of those having a difference of less than or equal to a threshold, A program for causing a computer to function as a granting device.   29. The content meta according to claim 28, wherein the certain threshold is an average of significance in each meta item of the browse history holding unit and an average of differences in significance of content in the meta item. A program for causing a computer to function as an information providing device.   The dynamic content meta-information adding unit, when there is some other content whose similarity with the content is equal to or greater than a threshold value for the certain content, assigns the meta-information and weight of the other certain content to the content meta The value obtained by multiplying the similarity and the acquired weight obtained from the information holding unit is added to the weight of the meta information of the certain content, and the meta information possessed by the other certain content is added to the certain content. 30. When there is no content, the value multiplied by the meta information is added to the certain content, so that the computer functions as the content meta information adding device according to any one of claims 24 to 29. Program for.   An inter-content similarity calculating unit that calculates the similarity between contents held by the inter-content similarity holding unit, and a content meta that is shared by the content according to the obtained similarity and is held in the content meta-information holding unit The program for functioning a computer as a content meta-information provision apparatus of any one of Claims 24 thru | or 30 further provided with an information provision part.   25. The content meta information adding device according to claim 24, wherein when the search condition is given, the content meta information held in the content meta information holding unit is searched using the search condition as a key, and the search A search unit for extracting content meta information whose value matches the condition and meta item value, and when the search condition and search meta information are given, the search condition and the search meta information are passed to the dynamic content meta information adding device. A series of processes for updating the content meta information in the content meta information holding unit, and then acquiring content meta information whose meta condition value matches the search condition from the content meta information holding unit using the search unit A program for causing a computer to function as a content search apparatus comprising a control unit that performs the above. The dynamic content meta information providing unit further includes:
When a predetermined past browsing situation of content is given, a past history corresponding to the browsing situation is read from the browsing history information holding unit, and the browsing history information is read based on the read browsing history information read. The degree of similarity between the contents in the situation is calculated, and the contents meta information held in the temporary contents meta information holding unit between the contents determined to be similar to each other based on the degree of similarity A process for giving the content meta information of the other party or increasing the weight of the content meta information corresponding to each content according to the similarity,
The search unit excludes the updated content from the search target from the content meta information holding unit, and sets the temporary content meta information holding unit as a search target using the search condition and the search meta information as a key. A program for causing a computer to function as the content search device according to claim 32.   The search unit includes contents having the same search condition as the given search condition and related meta information for the set of content and meta information in the content meta information holding unit and the temporary content meta information holding unit. 34. A program for causing a computer to function as a content search apparatus according to claim 33, wherein content having the same is extracted, and the content is sorted using the weight of meta information whose meta item value matches the search condition. .   It further comprises an input unit for acquiring a search condition for searching for content and passing the acquired search condition to the control unit, and an output unit for displaying the search result received from the control unit. A program for causing a computer to function as the content search device according to any one of claims 32 to 34.

Images