Tagging support device, program, and content search device

Abstract

To provide a tag provision support device, a program, and a content search device that can add a tag suitable for a content through simpler processing.SOLUTION: A tag provision support server 1 includes: a character string extraction processing unit 12 that extracts some character strings in a content having a text from the content; a tag provision unit 16 that tags the content with the character strings extracted by the character string extraction processing unit 12; and a tagged content registration unit 17 that causes a search DB 3 to store the tagged content tagged by the tag provision unit 16.SELECTED DRAWING: Figure 2

Description

Technical Field

[0001] The present invention relates to a tagging support device, a program, and a content search device.

Background Art

[0002] Conventionally, some contents on SNS (Social Networking Service) etc. have hash tags added by the poster. A hash tag is a search tag for searching contents, and is obtained by attaching a hash symbol (#) to the beginning of a word or phrase. By attaching a hash tag, it is possible to display a list of posts of other posters with the same hash tag, and it is possible to efficiently browse posts of users with a specific topic or the same interests. While hash tags can be freely added by the poster, there is a problem of lack of consistency.

[0003] Under such circumstances, for example, "an optimal tag proposal device that proposes tags to be inserted into the content transmitted by a transmitting device, collects the tags inserted into the content for each content in which a specific word appears, clusters the tags collected from the content in which the specific word appears for each specific word, determines a representative tag for each cluster from the tags belonging to the cluster according to the number of collections of each tag, when receiving a tag insertion request for the content to be transmitted from the transmitting device, collects the cluster of the specific word each time the specific word appears in the content to be transmitted, determines a candidate tag from the representative tags of the collected clusters according to the number of collections of each cluster, and returns it to the transmitting device, and when clustering, network the tags inserted into the same content by connecting them to each other, and each network is used as a cluster, the optimal tag proposal device." is disclosed (for example, Patent Document 1).

Prior Art Documents

Patent Documents

[0004] [Patent Document 1] Patent No. 5252593 [Overview of the project] [Problems that the invention aims to solve]

[0005] The technology described in Patent Document 1 above involves pre-registering words that frequently appear in content into a word database, and for each word registered in the word database, collecting tags already inserted into the content to create an optimal tag list database. Then, as tags for new content, representative tags from the optimal tag list database are returned as candidate tags based on the words contained in the content and the optimal tag list database. Therefore, it relied on using tags from existing content, and assumed that existing content was already tagged. Furthermore, the process of creating the optimal tag list database was cumbersome.

[0006] Therefore, the present invention aims to provide a tagging support device, a program, and a content search device that can assign tags appropriate to content through simpler processing. [Means for solving the problem]

[0007] The present invention solves the above problem by the following means. The first invention is a tagging support device comprising: a string extraction means for extracting a portion of a string from content having text; a tagging means for tagging the content with the string extracted by the string extraction means; and a content registration means for storing the tagged content tagged by the tagging means in a search database. The second invention is a tagging support device of the first invention, further comprising a keyword storage unit that stores keywords, wherein the string extraction means extracts the keyword as a string when the keyword stored in the keyword storage unit is included in the content. The third invention is a tagging support device according to the first or second invention, comprising a noun extraction means for analyzing the text of the content to extract nouns, wherein the string extraction means extracts predetermined nouns from the nouns extracted by the noun extraction means as strings. The fourth invention is a tagging support device of the third invention, wherein the string extraction means extracts, as a string, a predetermined noun from among the nouns extracted by the noun extraction means, the noun having an occurrence count of a threshold or more in the text. The fifth invention is a tagging support device of the third invention, wherein the string extraction means uses the tagged content registered in the search database to assign an importance score to the nouns extracted by the noun extraction means, and extracts the predetermined nouns having a score above a threshold as the string. The sixth invention is a tagging support device according to any of the first to fifth inventions, wherein the string extraction means extracts the string which is a proper noun from the content using a learning model that extracts proper nouns from the context of the text. The seventh invention is a tagging support device of any of the first to sixth inventions, comprising a related word acquisition means for acquiring related words related to the string extracted by the string extraction means from a related word storage unit that stores a plurality of related words, and the tagging means further tags the content with the related words acquired by the related word acquisition means. The eighth invention is a tagging support device that, in any of the first to seventh inventions, comprises a string output means for outputting the string extracted by the string extraction means, and a designated string receiving means for receiving a designated string specified by a user from among the strings output by the string output means, wherein the tagging means tags the content with the designated string received by the designated string receiving means. The ninth invention is a program for causing a computer to function as a tagging support device for any of the first to eighth inventions. The tenth invention is a content search device that is communicatively connected to a search database in which tagged content tagged by any of the tagging support devices from the first to the eighth inventions is stored, and comprises: a search term receiving means for receiving a search term; a search related term acquisition means for acquiring related terms related to the search term received by the search term receiving means from a related term storage unit that stores a plurality of related terms; a content extraction means for extracting tagged content having tags that match either the search term received by the search term receiving means or the related terms acquired by the search related term acquisition means from the search database; and a tag list output means for generating and outputting a tag list for each tag based on the tagged content extracted by the content extraction means. The eleventh invention is a content search device of the tenth invention, comprising: a tag selection receiving means that receives the selection of one of the tags from the tag list output by the tag list output means; and a content list output means that outputs a list of tagged content to which the tag received by the tag selection receiving means has been assigned, wherein the content list output means outputs a list of tagged content including the tag assigned to each tagged content. [Effects of the Invention]

[0008] According to the present invention, it is possible to provide a tagging support device, a program, and a content search device that can assign tags appropriate to content through simpler processing. [Brief explanation of the drawing]

[0009] [Figure 1] This diagram shows the overall configuration of the content search support system according to this embodiment. [Figure 2] This diagram shows the functional blocks of the tagging support server and the content search server according to this embodiment. [Figure 3] This figure shows an example of content to which tags are assigned by the tagging support server according to this embodiment. [Figure 4] This flowchart shows the content tagging process of the tagging support server according to this embodiment. [Figure 5] This flowchart shows the content search process of the content search server according to this embodiment. [Figure 6] This figure shows an example of the display on a terminal according to this embodiment. [Figure 7] This figure shows an example of the display on a terminal according to this embodiment. [Modes for carrying out the invention]

[0010] The following describes embodiments for carrying out the present invention with reference to the figures. However, this is merely an example, and the technical scope of the present invention is not limited thereto. (Embodiment) <Overall Configuration of Content Search Support System 100> Figure 1 shows the overall configuration of the content search support system 100 according to this embodiment. Figure 2 shows the functional blocks of the tagging support server 1 and the content search server 5 according to this embodiment.

[0011] The content search support system 100 shown in Figure 1 uses a tagging support server 1 (tagging support device) to tag content that does not have tags stored in the search database (DB) 3. The content search support system 100 also uses the tagging support server 1 to tag content received from terminal 7 and register it in the search DB 3. Furthermore, when the content search server 5 (content search device) receives a search term from terminal 8, the content search support system 100 sends a list of content that has the search term as a tag to terminal 8.

[0012] The content search support system 100 includes a tagging support server 1, a search DB 3, an associative word dictionary 4 (associative word memory unit), a content search server 5, a terminal 7, and a terminal 8, which are each connected via a communication network N. The communication network N is a network between the tagging support server 1, the search DB 3, the associative word dictionary 4, the content search server 5, the terminal 7, and the terminal 8, and is, for example, a communication network such as an Internet line. Also, the communication network N may be wired or wireless.

[0013] <Tagging Support Server 1> The tagging support server 1 is a server that performs a process of attaching tags suitable for the content to the content. FIG. 2(A) shows the functional blocks of the tagging support server 1. As shown in FIG. 2(A), the tagging support server 1 includes a control unit 10, a storage unit 20, and a communication interface unit 29.

[0014] The control unit 10 is a central processing unit (CPU) that controls the entire tagging support server 1. The control unit 10 appropriately reads and executes the operating system (OS) and various application programs stored in the storage unit 20, and cooperates with the above-described hardware to execute various functions. The control unit 10 includes a content acquisition unit 11, a character string extraction processing unit 12 (noun extraction means, character string extraction means), an associative word acquisition unit 13 (associative word acquisition means), a character string output unit 14 (character string output means), a designated character string reception unit 15 (designated character string reception means), a tagging unit 16 (tagging means), and a tagged content registration unit 17 (content registration means).

[0015] The content acquisition unit 11 acquires the content to which tags are to be attached. More specifically, the content acquisition unit 11 acquires the content to which tags are to be attached by extracting the content without tags from the search DB 3. Also, the content acquisition unit 11 receives the content to be tagged from the terminal 7. Here, the content acquired by the content acquisition unit 11 is, for example, a web page. And the web page contains text written in a markup language, such as HTML (Hypertext Markup Language).

[0016] The string extraction processing unit 12 extracts strings from the content acquired by the content acquisition unit 11. Here, the strings extracted by the string extraction processing unit 12 are a portion of the text content, such as keywords or key phrases that succinctly represent the content. The string extraction processing unit 12 may extract a keyword as a string if the keyword stored in the keyword dictionary 22 (keyword storage unit) (described later) is included in the content. Furthermore, the string extraction processing unit 12 may analyze the text of the content to extract nouns, and then extract a predetermined noun from the extracted nouns as a string. Here, a predetermined noun refers to, for example, a compound noun or a proper noun. Furthermore, the string extraction processing unit 12 may analyze the text of the content to extract nouns, and then extract as strings the nouns that are present in the text in quantities equal to or greater than a threshold.

[0017] Furthermore, the string extraction processing unit 12 may use tagged content registered in the search DB3 to assign importance scores to the extracted nouns, and extract nouns with scores above a threshold as strings. Here, the assignment of importance scores can be done, for example, using TF-IDF. Furthermore, the string extraction processing unit 12 may use a learning model stored in the learning model storage unit 23, which extracts proper nouns from the context of text, to extract proper noun strings from the content. The related word acquisition unit 13 obtains related words from the related word dictionary 4 that are related to the string extracted by the string extraction processing unit 12. Here, related words include synonyms, synonyms, etc.

[0018] The string output unit 14 outputs the string extracted by the string extraction processing unit 12 to the terminal 7. The string output unit 14 also outputs the related words acquired by the related word acquisition unit 13 to the terminal 7. The specified string receiving unit 15 receives the specified string from the terminal 7, which is selected from the strings and related words output by the string output unit 14 and specified by the user of the terminal 7. The processing performed by the string output unit 14 and the specified string receiving unit 15 is optional. For example, if the content acquisition unit 11 acquires content from the terminal 7, the control unit 10 may perform the processing performed by the string output unit 14 and the specified string receiving unit 15. On the other hand, if the content acquisition unit 11 acquires content from the search DB 3, the control unit 10 may choose not to perform the processing performed by the string output unit 14 and the specified string receiving unit 15.

[0019] The tagging unit 16 attaches the string extracted by the string extraction processing unit 12 to the corresponding content. The tagging unit 16 also attaches the related words obtained by the related word acquisition unit 13 to the corresponding content. Furthermore, if processing has been performed by the string output unit 14, the tagging unit 16 attaches the specified string received by the specified string reception unit 15 to the corresponding content. The tagged content registration unit 17 registers the tagged content, to which the tagging unit 16 has assigned tags, in the search DB3. Here, if the tagged content registration unit 17 is already stored in the search DB3, it may update the content stored in the search DB3 with the tagged content. Details of each of these functions will be described later.

[0020] The storage unit 20 is a storage device such as a hard disk or semiconductor memory element for storing programs, data, etc., necessary for the operation of the tag assignment support server 1. The memory unit 20 comprises a program memory unit 21, a keyword dictionary 22, and a learning model memory unit 23. The program storage unit 21 is a storage area that stores various programs necessary for the tag assignment support server 1 to function. The program storage unit 21 stores the tag assignment support program 21a. The tag assignment support program 21a is a program for executing each of the functions of the control unit 10 described above.

[0021] The keyword dictionary 22 is a memory area that stores keywords. The keyword dictionary 22 pre-stores multiple keywords (words) to be used as tags. Note that the keyword dictionary 22 may also store key phrases in addition to keywords. The learning model memory unit 23 is a memory area for storing various learning models. For example, the learning model memory unit 23 may store a learning model that extracts proper nouns from the context of text. The communication interface unit 29 is an interface for communicating with the search DB 3, the related word dictionary 4, and the terminal 7.

[0022] <Searchable DB3> The search DB3 is a database that stores content. Before processing by the tagging support server 1 of the content search support system 100, the search DB3 may contain content that has not been tagged. Content stored in the search DB3 that has not been tagged is tagged by the tagging support server 1 and stored in the search DB3 as tagged content. Furthermore, the search database 3 is used when the content search server 5 searches for content. <Related Word Dictionary 4> Related word dictionary 4 is, for example, a dictionary listing common related words. Related word dictionary 4 stores multiple related words that have the same or similar meanings in association with each other. Related word dictionary 4 may also be, for example, a thesaurus.

[0023] <Content Search Server 5> The content search server 5 is a server that searches for and extracts content that has tags based on search terms from the search DB3, which stores tagged content. Figure 2(B) shows the functional blocks of the content search server 5. The content search server 5 shown in Figure 2(B) comprises a control unit 50, a storage unit 60, and a communication interface unit 69.

[0024] The control unit 50 is a CPU that controls the entire content search server 5. The control unit 50 works in cooperation with the aforementioned hardware to perform various functions by appropriately reading and executing the OS and various application programs stored in the memory unit 60. The control unit 50 includes a search term receiving unit 51 (search term receiving means), a search related term acquisition unit 52 (search related term acquisition means), a content extraction unit 53 (content extraction means), a list generation processing unit 54 (tag list output means), and a tag selection processing unit 55 (tag selection receiving means, content list output means).

[0025] The search term receiving unit 51 receives search terms from the terminal 8. The search-related term acquisition unit 52 acquires related terms from the related term dictionary 4 that are related to the search term received by the search term reception unit 51. The content extraction unit 53 extracts tagged content from the search DB 3 that has tags matching either the search term received by the search term receiving unit 51 or the related term obtained by the search related term acquisition unit 52.

[0026] The list generation processing unit 54 generates a tag list for each tag based on the tagged content extracted by the content extraction unit 53 and outputs it to the terminal 8. The tag selection processing unit 55 accepts the selection of a tag from the tag list, generates a list of tagged content to which the accepted tag is assigned, and outputs it to the terminal 8. Details of each of these functions will be described later.

[0027] The storage unit 60 is a storage device such as a hard disk or semiconductor memory element for storing programs, data, etc., necessary for the operation of the content search server 5. The storage unit 60 includes a program storage unit 61. The program storage unit 61 is a storage area that stores various programs necessary for the content search server 5 to function. The program storage unit 61 stores the content search program 61a. The content search program 61a is a program for executing each of the functions of the control unit 50 described above. The communication interface unit 69 is an interface for communicating with the search DB3, the related word dictionary 4, and the terminal 8.

[0028] Furthermore, a computer refers to an information processing device equipped with a control unit, memory device, etc., and both the tagging support server 1 and the content search server 5 are information processing devices equipped with a control unit, memory device, etc., and are therefore included in the concept of a computer. Furthermore, the tagging support server 1 and the content search server 5 are not limited to a single server each. They may consist of multiple servers, and there is no limit to the number of hardware components that make up the tagging support server 1 and the content search server 5.

[0029] For example, the hardware for the tagging support server 1 and the content search server 5 may include various servers such as a web server, a DB (database) server, and an application server as needed, and may be configured on a single server or on separate servers. Furthermore, the tagging support server 1 and the content search server 5 may be, for example, cloud-based. Furthermore, this may be implemented by a single server that has the functions of both a tagging support server 1 and a content search server 5.

[0030] <Terminals 7, 8> Terminal 7 shown in Figure 1 is, for example, a terminal used by users who create content or users who tag content. Terminal 8 is, for example, a terminal used by a user who wants to search for content stored in the search database 3. Terminals 7 and 8 can each be composed of, for example, a personal computer (PC) or a tablet device. Although not shown in the diagram, terminals 7 and 8 each include a control unit, a storage unit, a display unit, an input unit, a communication interface unit, and so on. Note that terminals 7 and 8 do not need to be separate devices; both functions can be performed on the same device.

[0031] <Content Description> Next, we will explain examples of content handled by the content search support system 100. Figure 3 shows an example of content 30 to which tags are assigned by the tagging support server 1 according to this embodiment. The content 30 shown in Figure 3 is a webpage about a company's products. Content 30 provides a description of the products in text format. Content 30 contains various information about the product. In particular, the "Overview" and "Features" sections contain information that concisely summarizes the content of Content 30. In the following examples, the content 30 stored in the search DB3 will be described as a webpage related to the products of a company that has implemented the content search support system 100, as illustrated in Figure 3. Furthermore, when performing tagging, the processing will be performed on the "Overview" and "Features" sections of the content 30.

[0032] <Explanation of the process> Next, we will explain the processing performed by the tag assignment support server 1. Figure 4 is a flowchart showing the content tagging process of the tagging support server 1 according to this embodiment. As a prerequisite for this process, the search database 3 must contain at least untagged content. The search database 3 may also contain tagged content.

[0033] In step S (hereinafter simply referred to as "S") 11 of Figure 4, the control unit 10 (content acquisition unit 11) extracts content that has not been tagged from the search DB 3. If multiple pieces of content that have not been tagged are extracted, the control unit 10 performs the processing from S12 onwards for each piece of content. In S12, the control unit 10 (string extraction processing unit 12) performs a string extraction process to extract some strings from the content. Several methods are possible for extracting some strings from the content, and any of these methods may be used.

[0034] Here, the process of extracting strings will be explained based on an example. (Example 1) Example 1 uses the keyword dictionary 22. The control unit 10 checks whether the keywords registered in the keyword dictionary 22 are included in the content, by referring to the keyword dictionary 22 for each keyword. The control unit 10 then obtains keywords from the keyword dictionary 22 that are included in the content, as strings to be used for tagging. For example, if the keywords "microwave oven," "heat treatment," "heat-resistant specifications," and "pouch" are registered in keyword dictionary 22, then all of the above keywords will be retrieved in content 30 as shown in Figure 3.

[0035] (Example 2) Example 2 involves extracting nouns from content and then extracting specific nouns as keywords. The control unit 10 performs, for example, morphological analysis on the content text to extract nouns. The control unit 10 then obtains, for example, compound nouns and proper nouns from the extracted nouns as strings for tagging. Alternatively, the control unit 10 may obtain, from the extracted nouns, frequently occurring nouns whose occurrence count is above a threshold as strings for tagging. In content 30 shown in Figure 3, compound nouns and proper nouns such as "microwave oven," "heat treatment," "heat-resistant specifications," and "○○ (proper noun)" are obtained as strings. In addition, in content 30, frequently occurring nouns such as "range" are obtained as strings.

[0036] (Example 3) Example 3 extracts nouns from the content and extracts keywords using statistical values. The control unit 10 uses the content registered in the search DB3 (whether tagged or not) to assign an importance score to the extracted nouns. This scoring can be done, for example, using TF-IDF. Here, TF-IDF is a value used when extracting characteristic words from a document. When there are several documents, it quantifies which words are important to a particular document based on the words that appear in them and their frequencies. TF (Term Frequency) represents the number of occurrences in a document; the more often a word appears in a document, the more likely it is to be important. IDF (Inversed Document Frequency) is the number of documents in which a word appears; a word that appears in many documents is less likely to be a characteristic word. The control unit 10 then extracts nouns that have a score equal to or greater than the threshold calculated using TF-IDF as strings.

[0037] (Example 4) Example 4 uses a learning model to extract proper nouns from the context of text. Here, the learning model can be trained using deep learning methods such as BERT (Bidirectional Encoder Representations from Transformers)-CRF (Conditional Random Fields).

[0038] The control unit 10 may use any of the methods described in the above embodiment, or a combination of methods, for extracting the string. The control unit 10 may, for example, not convert all of the keywords obtained in Example 1 into strings, but rather convert only those keywords that appear multiple times into strings.

[0039] In step S13 of Figure 4, the control unit 10 obtains related words from the related word dictionary 4 that are associated with the string extracted in the process of S12. For example, if the string "working from home" is extracted, the control unit 10 refers to the related word dictionary 4 and retrieves, for example, "telework" as a synonym and, for example, "online system" as a related word. In S14, the control unit 10 (tagging unit 16) assigns the string extracted in the S12 process and the related words obtained in the S13 process as tags to the content. In S15, the control unit 10 (tagged content registration unit 17) registers the tagged content with the tagged content by replacing the content extracted in the S11 process with the tagged content with the tagged content. After that, the control unit 10 terminates this process.

[0040] The above explanation described how tags are added to content already registered in the search database. When introducing this content search support system 100, this process allows tags to be added to content already stored in the search database that does not yet have tags.

[0041] Furthermore, for example, when registering content from terminal 7 to the search DB3, the tag assignment support server 1 can also perform content tag assignment processing. In this case, for example, instead of processing S11 in Figure 4, the control unit 10 (content acquisition unit 11) acquires the content from terminal 7. Then, for example, after processing S13, the control unit 10 (string output unit 14) may output the extracted string and related words to terminal 7 so that the user of terminal 7 can check the strings that are candidates for tags. Then, the control unit 10 (specified string reception unit 15) receives the specified string specified by the user from terminal 7, and in S14, the control unit 10 (tag assignment unit 16) assigns the specified string received from terminal 7 as a tag to the content. In this way, the tag assignment support server 1 can assign the tags specified by the user to the content from the tag candidates that the user has reviewed.

[0042] Next, we will explain the processing performed by content search server 5. Figure 5 is a flowchart showing the content search process of the content search server 5 according to this embodiment. Figures 6 and 7 show examples of displays on terminal 8 according to this embodiment. In S51 of Figure 5, the control unit 50 (search term receiving unit 51) receives a search term from the terminal 8 to search for content. The search term is a keyword specified by the user of the terminal 8.

[0043] In S52, the control unit 50 (search related word acquisition unit 52) ​​acquires related words related to the search term received in S51 from the related word dictionary 4. In S53, the control unit 50 (content extraction unit 53) extracts tagged content from the search DB3 that has tags matching either the search term received in S51 or the related term obtained in the processing of S52. In S54, the control unit 50 (list generation processing unit 54) generates a tag list based on the extracted tagged content, and outputs the generated tag list to the terminal 8.

[0044] In S55, the control unit 50 (tag selection processing unit 55) determines whether or not it has received a selection of a tag from the user's tag list from the terminal 8. If a selection of a tag has been received (S55: YES), the control unit 50 moves the process to S56. On the other hand, if a selection of a tag has not been received (S55: NO), the control unit 50 terminates this process. In S56, the control unit 50 (tag selection processing unit 55) generates a list of tagged content to which the received tags have been assigned, and outputs the generated list of content to the terminal 8. After that, the control unit 50 terminates this process.

[0045] Figure 6 shows an example of the tag list screen 80 output to terminal 8. When searching, the tag list screen 80 displays only the search criteria area 81. Then, on terminal 8, when the user specifies a search term in the search term input area 81a in the search criteria area 81 and selects the search button 81b, the tag list screen 80 changes to include the list area 82 below the search criteria area 81. The tag list screen 80 shown in Figure 6 is the search results screen when keyword search is specified as the condition in the search condition area 81 and "security" is specified in the search term input area 81a. The control unit 50 retrieves tagged content from the search DB3 that has tags containing either the search term "security" or related terms to "security". The control unit 50 then aggregates the number of tagged contents for each tag attached to the retrieved tagged content and generates and outputs a tag list including the tags and the aggregated count.

[0046] In the example of the tag list screen 80 shown in Figure 6, the list labeled "Topic" was created by the content search process described above. On the other hand, the lists labeled "Target Field" and "Technical Field" were created based on predetermined locations within the tagged content. Then, when the user performs an operation to select a topic (an operation to select a tag), the control unit 50 outputs the content list screen 90 shown in Figure 7 to the terminal 8.

[0047] The content list screen 90 includes an overview area 91 and displays an overview of each content item. The summary area 91 includes the tag area 91a. The tag area 91a is the area that outputs the tags assigned to the content output in the summary area 91. Each tag contains a string and a number in parentheses. The number in parentheses indicates the number of tagged content items stored in the search DB3 to which that string has been assigned as a tag. When a user selects a tag in the tag area 91a, the selected tag is output to the input area of ​​the search area 92 at the top of the content list screen 90, making it possible to further search for and output content that has that tag attached.

[0048] Thus, according to this embodiment, the content search support system 100 has the following effects. (1) The tagging support server 1 extracts some strings from the content, tags the content with the extracted strings, and stores the tagged content in the search DB 3. Therefore, the tag assignment support server 1 can assign tags appropriate to the content.

[0049] (2) The tagging support server 1 extracts keywords as strings when the keywords stored in the keyword dictionary 22 are included in the content. Therefore, since keywords stored in the keyword dictionary 22 can be retrieved as tags, appropriate tags can be assigned to content with simpler processing. In addition, since tags can be assigned that depend on the keyword dictionary 22, variations in the assigned tags can be reduced.

[0050] (3) The tagging support server 1 analyzes the text of the content to extract nouns, and then extracts specific nouns from the extracted nouns as strings. Furthermore, among the extracted nouns, those nouns whose occurrence count in the content exceeds a certain threshold are extracted as strings. Furthermore, using tagged content registered in the search DB3, a score indicating importance is assigned to the extracted nouns, and nouns with a score above a certain threshold are extracted as strings. Therefore, for example, compound nouns and proper nouns can be assigned as tags, and the assigned tags can represent the characteristics of the content. Furthermore, frequently occurring nouns can be assigned as tags, and these tags can represent the characteristics of the content. Furthermore, nouns with an importance score above a certain threshold can be assigned as tags, and these tags can represent the characteristics of the content.

[0051] (4) The tagging support server 1 uses a learning model stored in the learning model storage unit 23 to extract strings that are proper nouns from the content, which are proper nouns. Therefore, proper nouns extracted through simpler processing can be assigned as tags, and these tags can represent the characteristics of the content.

[0052] (5) The tagging support server 1 retrieves related words associated with the extracted string from the related word dictionary 4 and further tags the content with the retrieved related words. Therefore, the tags that can be attached to content can be expanded.

[0053] (6) The tagging support server 1 outputs the extracted strings, accepts the specified strings specified by the user from the output strings, and tags the content with the accepted specified strings. Therefore, it is possible to present users with a list of suggested tags to assign to their content. Then, the tags specified by the user from among the suggested tags can be assigned to the content. As a result, it is possible to avoid assigning tags that the user deems inappropriate.

[0054] (7) The content search server 5 retrieves related terms from the related term dictionary 4 that are associated with the received search term, extracts tagged content from the search DB 3 that has tags that match either the search term or the related terms, and generates and outputs a tag list for each tag based on the extracted tagged content. Therefore, content tagged by the tagging support server 1 can be searched using the tags. In this process, content tagged not only with tags matching the search term but also with tags matching related terms is extracted, thus broadening the range of content extracted by the search.

[0055] (8) The content search server 5 accepts the selection of a tag from the tag list and outputs a list of tagged content to which the accepted tag has been assigned. The content list then includes the tags assigned to the tagged content. Therefore, tags can be used to search for tagged content that has been assigned those tags.

[0056] Although embodiments of the present invention have been described above, the present invention is not limited to the embodiments described above. Furthermore, the effects described in the embodiments are merely a list of the most preferred effects arising from the present invention, and the effects of the present invention are not limited to those described in the embodiments. The embodiments described above and the modified forms described later can be used in combination as appropriate, but a detailed explanation is omitted.

[0057] (Transformed form) (1) In this embodiment, we have described an example in which a string to be assigned as a tag is extracted from a predetermined part of the content, but we are not limited to this. The target of string extraction may be another part, or it may be all of the content. However, it is more desirable to extract the string including the part in which the characteristics of the content are described.

[0058] (2) In this embodiment, the example described is tagging content that does not have tags assigned to it, but it is not limited to this. The content may already have tags assigned to it by other methods. In that case, the tags assigned by the tag assignment support server and the tags that have already been assigned may be managed separately. In this way, even content used in existing systems can be searched by processing by the content search server as a new search process. In particular, data migration processes can be easily carried out. [Explanation of Symbols]

[0059] 1. Tagging support server 3. Search Database 4. Related Word Dictionary 5. Content Search Server 7,8 terminals 10,50 Control Unit 11. Content Acquisition Section 12. String Extraction Processing Unit 13 Related Word Acquisition Section 14 String output section 15. String Reception Section 16. Tagging section 17. Tagged Content Registration Section 20,60 storage section 21a Tagging Support Program 22 Keyword Dictionary 23 Learning Model Memory Unit 30 Contents 51 Search term reception department 52 Search related term acquisition unit 53 Content Extraction Unit 54 List generation processing unit 55 Tag selection processing unit 61a Content Search Program 80 Tag List Screen 90 Content List Screen 100 Content Search Support System N Communication Network

Claims

1. A string extraction means for extracting a portion of a string from content that contains text, Related word acquisition means retrieves related words related to the string extracted by the string extraction means from a related word storage unit that stores a plurality of related words, A tagging means that tags the content with the string extracted by the string extraction means and the related words acquired by the related word acquisition means, Content registration means for storing the tagged content tagged by the aforementioned tagging means in a searchable database, A tagging support device equipped with the following features.

2. In the tagging support device according to claim 1, It is equipped with a keyword memory unit that stores keywords, The string extraction means is a tagging support device that extracts the keyword stored in the keyword storage unit as the string when the keyword is included in the content.

3. In the tagging support device according to claim 1, The system includes a noun extraction means for analyzing the text of the aforementioned content and extracting nouns, The string extraction means is a tagging support device that extracts a predetermined noun from the nouns extracted by the noun extraction means as the string.

4. In the tagging support device according to claim 3, The string extraction means is a tagging support device that extracts, as a string, predetermined nouns from among the nouns extracted by the noun extraction means, whose occurrence count in the text is equal to or greater than a threshold value.

5. In the tagging support device according to claim 3, The string extraction means is a tagging support device that uses the tagged content registered in the search database to assign an importance score to the nouns extracted by the noun extraction means, and extracts the predetermined nouns having a score above a threshold as the string.

6. In the tagging support device according to claim 1, The string extraction means is a tagging support device that extracts the string, which is a proper noun, from the content using a learning model that extracts proper nouns from the context of text.

7. In the tagging support device according to claim 1, A string output means that outputs the string extracted by the string extraction means, A specified string receiving means that receives a specified string specified by the user from among the strings output by the string output means, Equipped with, The tagging means is a tagging support device that further tags the content with the specified string received by the specified string receiving means.

8. Computers, A string extraction means for extracting a portion of a string from content that contains text, Related word acquisition means retrieves related words related to the string extracted by the string extraction means from a related word storage unit that stores a plurality of related words, A tagging means that tags the content with the string extracted by the string extraction means and the related words acquired by the related word acquisition means, Content registration means for storing the tagged content tagged by the aforementioned tagging means in a searchable database, A program designed to function as such.

9. A string extraction means for extracting a portion of a string from content having text, A tagging means for tagging the string extracted by the string extraction means to the content, Content registration means for storing the tagged content tagged by the aforementioned tagging means in a searchable database, A content search device that is communicatively connected to the search database in which the tagged content, which has been tagged by a tagging support device, is stored, A search term receiving method that accepts search terms, A search related word acquisition means retrieves related words related to the search term received by the search term receiving means from a related word storage unit that stores multiple related words, Content extraction means for extracting tagged content from the search database that has tags matching either the search term received by the search term receiving means or the related term obtained by the search related term acquisition means, A tag list output means generates and outputs a tag list for each tag based on the tagged content extracted by the content extraction means, A content search device equipped with the following features.

10. In the content search device according to claim 9, A tag selection receiving means that accepts the selection of one tag from the tag list output by the tag list output means, A content list output means outputs a list of tagged content to which the tag has been received by the tag selection receiving means, Equipped with, The content list output means is a content search device that outputs a list of tagged content, including the tags assigned to each tagged content.

Images