JP2007193697A - Information collection apparatus, information collection method and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To facilitate a reliable search through posted information by avoiding the generation of unintended search results due to homonyms and the like. <P>SOLUTION: A metadata assignment support device 230 is connected to a communication network 200, and comprises a word extraction part 304 for extracting words by natural language analysis from a document included in posted information, a vocabulary system holding part 306 for holding an ontology-based vocabulary system 354, a candidate output part 308 for outputting a keyword and candidates for attribute values of the keyword out of the words extracted by the word extraction part according to relations defined in the vocabulary system, a metadata generation part 310 for generating metadata 360 pairing the keyword, attributes, and words selected by a user from the candidates, and a metadata holding part 312 for holding the generated metadata. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to an information collection device, an information collection method, and a program for adding metadata to posted information. More specifically, even if there is a synonym or synonym, it is possible to detect related specific contents, and The present invention relates to an information collection apparatus, an information collection method, and a program capable of sharing information with search users in the same field.

  In the Internet typified by a computer communication network (communication network), the posting information (WebPage) generated on the assumption that each computer publishes through the WWW (World Wide Web) system corresponds to the posting information. It is possible to browse using a browser.

  When searching for desired information on the WWW, a full-text search technique is used that extracts posted information including a specific keyword input by the user from all posted information. Here, an index of keywords included in the posted information is created in advance, and a large amount of posted information can be searched at high speed by referring to the index.

  However, in the posting information generated by various people from various viewpoints, the terms used are not uniform, and the same thing is expressed in different terms (synonyms) or different things are expressed in the same terms (synonyms). It is often generated by expressing or changing the presence or absence of long sounds such as “computer” and “computer”, or changing katakana hiragana / alphabet letters (notation). Due to such a difference in notation, there are cases where desired posting information cannot be properly searched or posting information that is not intended is extracted.

  In order to avoid such an erroneous search, in a closed system such as a literature search, it is also considered to execute a search by assigning a control keyword. A technique is also known in which all keyword candidates are automatically extracted from documents to be searched, and corresponding documents are searched from posted information according to the keywords. However, even in such a technique, an excessive search result is generated because a keyword is searched with an unintended meaning such as a synonym, or an unnecessary keyword is searched.

As a method to avoid the occurrence of the above excessive search results, the user is sequentially asked which meaning the user wants to search for the synonyms, and the document is searched after confirming whether keywords are necessary. A technique is known (for example, Patent Document 1). There is also known a technique for quickly displaying a keyword search result by associating information specifying a bulletin information with a keyword to bulletin information published by a user using a specific system (for example, patents). Reference 2).
JP-A-8-314974 JP 2005-202895 A

  However, although the technology that asks the user the meaning of a keyword has a high level of exclusivity against the keyword's synonyms, it has not been established, so it supports a search with a high degree of freedom based on the keyword's attributes. I can't.

  Further, in the technology for associating the specific information of the posted information with the keyword, an unintended link is generated for the above-mentioned synonym and there is a possibility that a so-called “false explosion” may occur. In addition, since it is specified by a single keyword, there is no special function for adjusting the search results when there are too few or too many search results.

  In addition, the above technique cannot sufficiently cope with the case of synonyms. For example, when the user inputs “event” at the time of search, various information such as “host organization”, “entrance fee”, “participation qualification”, and the like are output as a result. Search results are often unnecessary. Therefore, even when searching from synonyms, it is desired that only information corresponding to user needs is obtained as a search result.

  The present invention has been made in view of the above-described problems of conventional search systems, and an object of the present invention is to avoid the generation of unintended search results by means of synonyms, etc., and to search posted information easily and reliably. To provide a new and improved information collecting apparatus, information collecting method and program.

  In order to solve the above-described problems, according to an aspect of the present invention, a word extraction unit connected to a communication network and extracts a word from a document included in bulletin information by natural language analysis; a vocabulary based on ontology A lexical system holding unit that holds the system; a candidate output unit that outputs a keyword and a word candidate extracted by the word extraction unit that should be an attribute value of the keyword in accordance with the association defined in the vocabulary system A metadata generation unit that generates metadata in which a word selected by the user from the keyword, attribute, and candidate is paired; and a metadata holding unit that holds the generated metadata. An information collecting device is provided.

  In order to improve the search accuracy of bulletin information including documents, words, and images (still images, moving images), in the present invention, metadata that abstractly describes the description content of the bulletin information is included in the public bulletin information. Give. Metadata includes not only a keyword representing the posting information but also an attribute, an attribute value, and other related information for the keyword. The posted information can be further characterized by such metadata, and the search target can be easily narrowed down.

  Also, by extracting the metadata attribute values from the posting information together with the keywords, it is possible to save the user from having to manually input the attribute values again. The metadata can be generated only by the process of selecting.

  A field selection unit that allows a user to select a desired field is further included. The lexical system holding unit includes a plurality of vocabulary systems provided for each field, and the candidate output unit is selected by the user by the field selection unit. Attribute value candidates may be output according to the vocabulary system of the field.

  The above vocabulary system may have different associations for each field or category. In such a case, more accurate metadata can be generated by applying a lexical system for each field. In such a configuration, the user who gives the metadata can select a desired field and use the vocabulary system of the field to extract vocabulary specialized for the field and generate metadata suitable for search. Is possible.

  A desired field estimation unit for estimating a field desired by the user from an operation history of the user to the information collecting apparatus, and the vocabulary system holding unit includes a plurality of vocabulary systems provided for each field, and the candidate output The section may output candidate attribute values according to the estimated lexical system of the field.

  In such a configuration, the user's operation history of the information collecting device is maintained, and the target posted information field is estimated from the field of the posted information having a large number of accesses or the field having the specified number of times, and the estimated By using the vocabulary system of a field, it is possible to extract vocabulary specialized for that field and generate metadata suitable for search.

  A posting field estimation unit for estimating a field to which the posting information belongs based on the word extracted by the word extraction unit; and the vocabulary system holding unit includes a plurality of vocabulary systems provided for each field, The candidate output unit may output attribute value candidates according to the estimated lexical system of the field.

  In such a configuration, a vocabulary specialized for a particular field is estimated by automatically estimating the field of the posted information from one or more words extracted from the bulletin information and using the vocabulary system of the estimated field. Can be extracted, and metadata suitable for search can be generated.

  A condition holding unit for holding metadata conditions desired by the user; and an additional notification unit for notifying the user when metadata satisfying the above conditions is added to the metadata holding unit by another user; May be further provided.

  When the user adds (registers) the specified field or metadata by specifying the desired field, metadata, metadata conditional expression, etc., the fact that it has been added, and / or Alternatively, the added information itself can be obtained. With this configuration, each user takes the initiative to use the information collection device in order to receive information from other users and to provide information to other users. Granting is promoted.

  The above information collection device is represented by an aggregate of a plurality of components, but is not limited to the case where each component or functional module belongs to a single device, and each component is a separate device or a separate device. It may be formed by being incorporated in the apparatus.

  In order to solve the above-described problem, according to another aspect of the present invention, a word extraction step of extracting a word by a natural language analysis from a document included in posted information; according to an association defined in a vocabulary system based on an ontology, A candidate output step for outputting a keyword and a word candidate extracted by the word extraction step to be an attribute value of the keyword; a meta-data pairing the keyword, the attribute, and a word selected by the user from the candidate There is provided an information collecting method comprising: a metadata generating step for generating data; and a metadata holding step for holding the generated metadata.

  Also provided are a program for causing a computer to perform the information collecting method and a storage medium storing the program.

  The constituent elements corresponding to the dependent claims in the information collecting apparatus and the explanation thereof can be applied to the information collecting method, program, and storage medium.

  As described above, according to the present invention, it is possible to search bulletin information easily and reliably while avoiding generation of an unintended search result by a synonym or the like by making metadata attached to the bulletin information a search target. It becomes possible to do.

  In addition, bulletin information that is assumed to be read by humans is not suitable for machine processing, but since metadata corresponding to machine processing is added to such bulletin information, it is linked with applications that use metadata. Becomes easier. For example, since information such as the date of an event can be extracted using metadata, it is possible to perform processing such as listing only information related to events held during a specific period.

  Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the present specification and drawings, components having substantially the same functional configuration are denoted by the same reference numerals, and redundant description is omitted.

  Currently, it is possible to search for desired posting information using a computer connected to a communication network such as the Internet. However, since the Web is an open system, there is a problem that it is impossible to detect appropriate posted information because there are too many corresponding documents in a search using only simple keywords.

  In order to improve the search accuracy of the bulletin information, it may be possible to attach a single or plural control keywords to the entire bulletin information document. In this embodiment, metadata is used as a standard specification that characterizes the description content of the bulletin information. Is granted. Metadata includes not only a keyword specifying the posting information but also an attribute, an attribute value, and other related information for the keyword. The posted information can be further characterized by such metadata, and the search target can be easily narrowed down.

  The above metadata is generated as a predicate and object when the keyword representing the posted information is the subject. Here, the part corresponding to the subject is the type of the keyword, and the part corresponding to the predicate is the attribute corresponding to any attribute or object of the type. Is an arbitrary attribute value in the attribute. Here, this [type]-[attribute]-[attribute value] relationship is called ontology, and based on this definition, each [attribute] and [attribute value] for a specific [type] is defined. Things are called vocabulary systems.

  FIG. 1 is an explanatory diagram showing an example of the vocabulary system described above. Such a vocabulary system is expressed by a hierarchy. For example, for “type” of “event” 100, “attribute” of “event” 100, “date of holding” 110, “place of holding” 112, “event name” 114, “organizer” 116 and the like are defined, and each [attribute] has a specific [attribute value] “date, period” 120, “place name, place expression” 122, Each category includes “title expression” 124, “person name, organization name” 126, and the like.

  FIG. 2 is a text diagram showing an example of the document in the posted information. When metadata is generated from such a document, for example, “Ubiquitous Exhibition” — “Date” — “2006” according to the association of “Event” 100— “Date” 110— “Date, Period” 120 of the vocabulary system. "January 1 to January 10" can be set as metadata.

  FIG. 3 is an explanatory diagram showing an example of metadata set based on the vocabulary system shown in FIG. The user attaches such metadata to the bulletin information, represents the bulletin information, and supports searches by other users. Such metadata is generated corresponding to the vocabulary system of FIG. 1 and includes the above-mentioned metadata “Ubiquitous Exhibition”-“Date”-“January 1 to January 10, 2006”. . In FIG. 3, four metadata are given as an example. However, the number of metadata is not limited to this, and various metadata composed of all terms can be generated.

  Such metadata is usually set manually by the user each time. However, under such a setting method, the load on the user becomes excessive and the above-mentioned system is not used much. Therefore, in this embodiment, the [attribute value] defined as metadata is extracted from the posting information together with the keyword, so that the user does not have to manually input the [attribute value] again and is automatically extracted. The metadata is generated only by the process of allowing the user to select [attribute value] from the candidate words.

(First embodiment: metadata assignment support apparatus)
FIG. 4 is a block diagram illustrating an outline of an information collection system in which the information collection apparatus is used. Such an information collection system includes a communication network 200, an information resource server 210, a terminal device 220, and a metadata addition support device 230 as an information collection device. In the information collection system, it is possible to provide an environment that facilitates user search processing, and also allows users to share information. In the following, the metadata providing device 230 is described as an information collecting device in the information collecting system. However, the metadata providing device 230 is used from a wide viewpoint such as information collection as well as metadata addition. Is possible.

  The communication network 200 is a bidirectional communication network such as the Internet, an intranet, or a LAN (Local Area Network), and is connected to the information resource server 210, the terminal device 220, the metadata assignment support device 230, and the like.

  The information resource server 210 has one or more posted information (WebPage), and provides the posted information via the communication network 200 in response to a user's browsing request.

  The terminal device 220 is usually managed by an individual user, searches for desired posting information by using a normal keyword or specific metadata, and accesses the metadata assignment support device 230 from the information resource server 210. Give metadata to the posted information provided.

  The metadata assignment support apparatus 230 adds meta information to the posted information acquired from the information resource server 210 in response to a request for metadata assignment from the information resource server 210, the terminal device 220, or an operator who operates the metadata assignment support apparatus 230. Give data. Further, in response to a posting information search command from the information resource server 210 and the terminal device 220, the held metadata is searched, and the posting information associated with the metadata is transmitted to the information resource server 210 and the terminal device 220. To do.

  As described above, all users of the information resource server 210, the terminal device 220, and the metadata assignment support apparatus 230 are targeted as users who use the metadata assignment support apparatus 230. For example, if it is an administrator of the information resource server 210, metadata is given to make the posting information available to many people, and if it is a user of the terminal device 220, other users who have the same hobby. The metadata is added to teach the posting information that deepens the interest of the user, and the operator of the metadata addition support device 230 provides the metadata so that more people can use the search engine.

  In addition, by associating product problems and solutions handled by electronic equipment manufacturer engineers and consulting service operators with metadata and publishing the posted information, other manufacturers' engineers and similar problems can be identified. It is possible to make it easier for a user to refer to the problem or solution. In this way, engineers can share problems and solutions, and even when users of a product do not know how to use the product, they can search for usage without going through the operator. Is possible.

  The information resource server 210, the terminal device 220, and the metadata assignment support device 230 may be configured by electronic devices such as a server, a personal computer, a PDA (Personal Digital Assistant), a mobile phone, and a game device.

  Next, the metadata assignment support apparatus 230 will be described in detail.

  FIG. 5 is a functional block diagram showing a schematic function of the metadata assignment support apparatus 230 in the first embodiment. The metadata assignment support apparatus 230 includes a bulletin information acquisition unit 300, a dictionary holding unit 302, a word extraction unit 304, a lexical system holding unit 306, a candidate output unit 308, a metadata generation unit 310, and metadata. And a holding unit 312.

  The bulletin information acquisition unit 300 acquires bulletin information 350 desired by the user via the communication network 200. The posting information 350 is formed in a markup language such as HTML (Hyper Text Markup Language) and is a target to which metadata is added.

  The dictionary holding unit 302 holds a dictionary 352 relating to word recognition such as morphemes, affixes, connection information, grammar information, and synonyms. According to the dictionary 352, the part of speech of the word extracted by the word extraction unit 304 described later and the meaning of the word can also be detected. For example, “Tokyo” can be determined as a place name, or “3000 yen” of “pay 3000 yen” can be determined as an amount of money from the relationship with the predicate “pay”.

  The word extraction unit 304 refers to the dictionary 352 held in the dictionary holding unit 302, deletes a tag unnecessary for specifying the document of the posting information 350, and converts the document included in the posting information 350 into a natural language. Analysis (syntactic analysis such as morpheme analysis and hook-and-loop analysis) breaks down into morphemes in which nouns and verbs are associated, and for example, removes characters that do not make sense by themselves, such as “Tenanoha” A keyword as a [type] representing the information 350 and a word that can represent an [attribute value] of the keyword (for example, a proper noun such as a person name or a place name) are extracted.

  Based on the ontology, the lexical system holding unit 306 determines the [type] of the keyword, what [attribute] the [type] takes, and what vocabulary can be taken as the [attribute value] in the [attribute] And a vocabulary system 354 that defines the specific value of, and its constraints.

  The vocabulary system 354 may be expressed in, for example, a description language such as W3C RDF (Resource Description Framework) or OWL (Web Ontology Language). Further, the vocabulary system 354 may be expressed by another ontology description language, or may be a table in which [type]-[attribute]-[attribute value] are simply made into a database. Since the general definition of the vocabulary system 354 has already been described with reference to FIG. 1, detailed description thereof is omitted here.

  Further, the restriction indicates a range or restriction that can be taken by the [attribute value], for example, a restriction of 15 years or older with respect to the [attribute] called “employee age”. In this way, even if the word extracted by the word extraction unit 304 has the expression “3 years old”, it is possible to exclude “3 years old” from the target of [attribute value] due to this restriction. In such a restriction, the [attribute value] can be narrowed as the range is limited, but there is a possibility that the intended [attribute value] cannot be detected.

  As another example, “3,000 yen” extracted from the document “personal computer discount of 3000 yen” is excluded because of the above restrictions because of the attribute “personal computer price”, which is extracted from the document “personal computer discount of 3000 yen”. The [Attribute value] of "PC price" is a word that contains only "yen", "\", and "$" in the word, and if a suitable word is not found, only a numerical value Are also subject to [Attribute Value].

  The candidate output unit 308 monitors the information resource server 210 or the terminal device 220 that the user has as a keyword candidate and the word candidate extracted by the word extraction unit 304, which should be the attribute value of the keyword, or The data is output (displayed) on the monitor 356 of the metadata assignment support apparatus 230.

  Here, the keyword as [type] representing the bulletin information is also output with candidates, but [type] may be the bulletin information itself. Thus, when [Type] is the bulletin information itself, it is not necessary to extract keyword candidates, and the output can be omitted.

  In addition, as the [attribute value] to be associated with each [attribute] of the keyword, a character string actually extracted by the word extraction unit 304 may be simply taken up, or is represented by a category that can be taken as [attribute value]. It is good. For example, for [attribute] “location” shown in FIG. 1, a specific “location name” or “location expression” that indirectly represents the location corresponds to the category of [attribute value]. . Therefore, as the [attribute value], it is possible to select “place name” such as “Chiyoda-ku ... address” in Tokyo, or “3-minute walk from Tokyo Station Yaesu Exit”.

  In addition, when the selection range is defined in advance as the [attribute value], for example, when the number of teams is limited as [attribute] “matching team”, the candidate output unit 308 displays the attribute value as [attribute value]. The entire team may be output.

  In addition, if the dictionary 352 includes a hierarchical concept (higher concept or lower concept) of the [attribute], a list of hierarchical concepts included in the dictionary 352 can be output as [attribute value]. For example, when an [attribute value] “TV” of a specific manufacturer is listed, the product name and product number of the “TV” of that manufacturer are further read out from the dictionary 352 and are used as metadata [attribute value] candidates. Is possible.

  Further, when there is a synonym in the above-mentioned candidate [type] or [attribute value], the word and the word representing the word may be output together. This representative will be described in detail later.

  If the candidate output unit 308 cannot find a candidate for [attribute value], conversely, each extracted word can be associated with [attribute]. For example, if there is no attribute value candidate of [attribute] “PC price”, the number “119,800” is extracted from the extracted word, and is replaced with “119,800 yen”, and [attribute] “PC” Attribute value.

  The metadata generation unit 310 sets the word selected by the user from the candidates output by the candidate output unit 308 as [attribute value], and combines the keyword 360 and [attribute] as [type] with the metadata 360. Is generated. A plurality of such metadata 360 may be given to one bulletin information.

  Here, when [attribute value] has another synonym and the other synonym is a representative term (representative), for example, a term according to JIS standard, the selected [attribute value] is represented by the representative The metadata 360 may be generated after the term is replaced. The above synonyms are, for example, the same word expressed in kanji, katakana, kana, romaji, or the presence or absence of a long sound. That is, “computer” is replaced with “computer” representing the representative, or “watch” is replaced with “clock” representing the representative.

  In addition, as described later, when the field to which the metadata 360 is assigned is specified, the extracted [attribute value] may be easily associated with the representative. When the word “CD” is extracted as the [attribute value], for example, it can be grasped as an abbreviation of “Compact Disc” or “Cash Dispenser”, but when the field is limited to music, “CD” is changed to “Compact”. It can be determined that it is synonymous with “Disc”, and if the field is limited to finance, “CD” can be determined as synonymous with “Cash Dispenser”.

  The metadata holding unit 312 holds the metadata 360 generated by the metadata generation unit 310. One or more metadata held in this way is referred to by the search engine, and the posted information with the corresponding metadata is displayed as the search result. The metadata is output in response to a request from another application. For example, an application for creating a calendar calls only metadata related to dates, and events related to metadata are displayed on the calendar on the corresponding days. Examples of the output format include RDF and OWL described above.

The dictionary holding unit 302, vocabulary holding unit 306, and metadata holding unit 312 are RAM, E ² PROM, nonvolatile RAM, flash memory, card memory, USB memory, HDD (Hard Disk Drive), and other storage media. The dictionary holding unit 302, the lexical system holding unit 306, and the metadata holding unit 312 may be configured as an integrated HDD.

  In addition, the bulletin information acquisition unit 300, the word extraction unit 304, the candidate output unit 308, the metadata generation unit 310, and the like may be formed as modules in the computer. In this case, the metadata addition support device 230 is configured. Each function may be performed by a central processing unit (CPU), a memory in which a program is stored, and each input / output circuit.

  As described above, the metadata 360 that characterizes the description content of the bulletin information 350 can be imparted to the public bulletin information 350 by the metadata grant support apparatus 230 described above. In the metadata 360, since [attribute] and [attribute value] are added to the keyword as [type] representing the posting information, the posting information 350 can be further characterized, and the search target can be easily narrowed down. Become. Further, since the [attribute value] is extracted from the bulletin information 350 itself, the metadata 360 can be generated only by a simple process of allowing the user to select the [attribute value]. A method for assigning metadata using the metadata assignment support apparatus 230 will be described below.

(Metadata grant support method)
FIG. 6 is a flowchart showing the processing flow of the metadata assignment support method according to the first embodiment. Hereinafter, the processing flow will be described together with the configuration of FIG.

  First, the bulletin information acquisition unit 300 of the metadata assignment support apparatus 230 acquires the bulletin information 350 by the HTTP protocol via the communication network 200 (S400). Then, the word extraction unit 304 uses the dictionary 352 to delete tags and the like that are unnecessary for specifying the document, decomposes the document included in the posting information 350 into morphemes by morphological analysis and syntax analysis, and displays the posting information 350 And a word that can be an [attribute value] of the keyword (for example, a proper noun such as a person name or a place name) is extracted (S402).

  Subsequently, the candidate output unit 308 follows the association of the vocabulary system 354 with the keyword candidates extracted in the word extraction step (S402) and the word extracted by the word extraction unit 304, which should be attribute values of the keywords. Candidates are listed, and a Web document representing the candidate list in a menu format is created (S404). Then, the Web document is output (displayed) to the monitor of the terminal of the user or the monitor 356 of the metadata assignment support apparatus 230 (S406). The display format at this time is not limited to the menu format, and various selection formats can be applied.

  Next, the metadata generation unit 310 sets the word selected by the user as [attribute value], and makes a pair with the keyword and [attribute], that is, a pair of [type]-[attribute]-[attribute value]. Then, metadata 360 is generated (S408). Finally, the metadata holding unit 312 stores the metadata 360 generated by the metadata generation unit 310 in a form that can be referred to by other terminal devices 220 (S410).

  Also provided are a program for causing a computer to perform the above-described metadata assignment support method, and a storage medium storing the program.

  According to the metadata grant support apparatus and metadata grant support method in the first embodiment, a document of bulletin information to be given metadata is analyzed, and not only a simple “keyword” but also its [attribute] and [attribute] Value] and allowing the user to select an appropriate [attribute value] to limit the meaning of the keyword or the posted information itself, even if the posted information includes a synonym. A more accurate search is possible without detection.

  Also, the posting information detection side can detect the intended posting information even if the keyword includes a synonym by specifying [attribute] and [attribute value] along with the keyword. Becomes easy.

  For example, if the user wants to search for what “events” are held at the Tokyo exhibition hall, he / she does not just enter “events” and keywords, but also “attributes” of “events” ([type]). Select “Venue” as “Enter”, and enter “Tokyo Exhibition Hall” as “Attribute Value”. Then, a search is performed with the metadata of “event” − “venue” − “Tokyo exhibition hall”, and “ubiquitous exhibition” as “event” can be easily obtained as a search result (see FIG. 3). .

  In addition, when the user searches for a document “currently held event”, the search engine analyzes the document, derives the current date “January 1” from the word “now”, "(Attribute)" is searched for "event" ([type]) including "January 1" in [attribute value]. As a result, an “Ubiquitous Exhibition” of “Event” whose “Date” is “January 1 to January 10, 2006” is obtained (see FIG. 3).

(Second Embodiment: Metadata Assignment Support Device)
In the first embodiment, the dictionary 352 and the vocabulary system 354 are represented as a single entity. However, depending on the field in which the posted information is used, the range of synonyms may be different, or the vocabulary concepts may be different. For example, the word “model” is used in the meaning of a model, model, etc. in the manufacturer, but is used in the fashion industry as a fashion model.

  In the second embodiment described below, when metadata is added to the posting information, the fields are distinguished according to the purpose of assignment, a dictionary 352 and a vocabulary system 354 for each field are selected, and such a purpose of assignment is selected. From this point of view, it is possible to generate metadata more easily and more suitable for the posted information by extracting words considered to be important in this field.

  The fields in the second embodiment include, for example, categories such as “sports”, “cooking”, and “school”, competitions such as “baseball” and “soccer”, and purposes such as “business” and “hobby”. , Can be distinguished in various ways.

  FIG. 7 is a functional block diagram showing a schematic function of the metadata assignment support apparatus 500 according to the second embodiment. The metadata assignment support apparatus 500 includes a bulletin information acquisition unit 300, a field selection unit 510, a dictionary holding unit 302, a word extraction unit 304, a vocabulary holding unit 306, a candidate output unit 308, and metadata generation. A unit 310 and a metadata holding unit 312 are included.

  The bulletin information acquisition unit 300, the word extraction unit 304, the candidate output unit 308, the metadata generation unit 310, and the metadata holding unit 312 that have already been described as components in the first embodiment substantially function. Therefore, the field selection unit 510, the dictionary holding unit 302, and the lexical system holding unit 306 having different configurations will be mainly described.

  The field selection unit 510 derives selectable fields in accordance with the bulletin information acquired by the bulletin information acquisition unit 300 and causes the user to select a desired field. It is also possible to allow the user to input a specific field without analyzing the bulletin information.

  The dictionary holding unit 302 holds a plurality of field-specific dictionaries 352 related to word recognition such as morphemes, affixes, connection information, grammatical information, and synonyms, and the word extraction unit 304 stores the fields selected by the field selection unit 510. With reference to the dictionary 352, a keyword representing the posting information 350 and a word that can represent an [attribute value] of the keyword (for example, a proper noun such as a person name or a place name) are extracted.

  The lexical system holding unit 306 holds a plurality of vocabulary systems 354 for each field, and the candidate output unit 308 follows keywords according to the vocabulary system 354 of the field selected by the field selection unit 510, and the [attribute value] of the keyword. The candidate words extracted by the word extraction unit 304 are output (displayed).

  The vocabulary system 354 may differ in vocabulary system for each field or category. In such a case, the metadata 360 with higher accuracy can be generated by applying the vocabulary system for each field. In such a configuration, the user who gives the metadata 360 selects a desired field, and by using the vocabulary system 354 of the field, the vocabulary specialized for the field can be extracted, and the metadata 360 suitable for the search can be obtained. Can be generated.

  Next, a configuration will be described in which the user's desired field is estimated based on the user's operation history instead of the user selecting.

  FIG. 8 is a functional block diagram showing a schematic function of the metadata assignment support apparatus 500 in another example of the second embodiment. The metadata assignment support apparatus 500 includes a bulletin information acquisition unit 300, a desired field estimation unit 520, a dictionary holding unit 302, a word extraction unit 304, a lexical system holding unit 306, a candidate output unit 308, metadata A generation unit 310 and a metadata holding unit 312 are included.

  The bulletin information acquisition unit 300, the word extraction unit 304, the candidate output unit 308, the metadata generation unit 310, and the metadata holding unit 312 that have already been described as components in the first embodiment substantially function. The description is omitted here, and here, the desired field estimation unit 520, the dictionary holding unit 302, and the lexical system holding unit 306 having different configurations will be mainly described.

  The desired field estimation unit 520 determines the user's operation history with respect to the metadata assignment support apparatus 500, for example, the history of bulletin information referred to in the past, the field of bulletin information with a large number of accesses, and the field with a large number of designated times. Estimate the field desired by the user.

  The dictionary holding unit 302 holds a plurality of dictionaries 352 for each field, and the word extraction unit 304 refers to the field dictionary 352 estimated by the desired field estimation unit 520 and uses the keyword representing the posting information 350 and the keyword Words that can represent [attribute value] (for example, proper nouns such as personal names and place names) are extracted.

  The lexical system holding unit 306 holds a plurality of vocabulary systems 354 for each field, and the candidate output unit 308 follows keywords and keyword [attribute values] according to the vocabulary system 354 of the field estimated by the desired field estimation unit 520. The candidate words extracted by the word extraction unit 304 are output (displayed).

  In such a configuration, the user's operation history of the metadata assignment support apparatus 500 is held, the field of the bulletin information 350 having a large number of accesses and the field having a large number of designated times are estimated, and the lexical system of the estimated field By using 354, a vocabulary specialized in the field can be extracted, and metadata 360 suitable for search can be generated.

  Next, a description will be given of a configuration in which the above-mentioned field is estimated on the side of the metadata providing support apparatus 500 based on the word extracted by the word extraction unit 304 instead of being estimated based on the user operation history.

  FIG. 9 is a functional block diagram showing a schematic function of the metadata assignment support apparatus 500 in another example of the second embodiment. The metadata assignment support apparatus 500 includes a bulletin information acquisition unit 300, a dictionary holding unit 302, a word extraction unit 304, a posting field estimation unit 530, a lexical system holding unit 306, a candidate output unit 308, a metadata A generation unit 310 and a metadata holding unit 312 are included.

  The bulletin information acquisition unit 300, the dictionary holding unit 302, the word extraction unit 304, the candidate output unit 308, the metadata generation unit 310, the metadata holding unit 312 described above as the constituent elements in the first embodiment Since the functions are substantially the same, repeated description is omitted, and here, the posting field estimation unit 530 and the lexical system holding unit 306 having different configurations will be mainly described.

  The posting field estimation unit 530 estimates a field to which the posting information belongs based on the word extracted by the word extraction unit 304.

  The lexical system holding unit 306 holds a plurality of vocabulary systems 354 for each field, and the candidate output unit 308 follows the vocabulary system 354 of the field estimated by the posting field estimation unit 530 and the keyword and [attribute value] of the keyword. The candidate words extracted by the word extraction unit 304 are output (displayed).

  In such a configuration, the field of the posted information is automatically estimated from one or more words extracted from the posted information, and the vocabulary system 354 of the estimated field is used to specialize in the field. Vocabulary can be extracted, and metadata 360 suitable for search can be generated.

  According to the second embodiment, by selecting a field for searching bulletin information, it is possible to reduce false detection by a synonym, so-called “misexplosion”, and to detect unnecessary candidates. Therefore, it becomes easier to detect desired bulletin information. Also, because synonym definitions and representative words are different for each field, even the same phrase in the original posted information is expressed as a different phrase at the metadata stage, and it is possible to avoid false explosions. .

(Third Embodiment: Metadata Assignment Support Device)
The metadata grant support apparatus according to the third embodiment gives an incentive to the grant of metadata 360 by the user, promotes the grant of metadata 360, and consequently increases the number of posting information to which the metadata 360 is given. Let In addition, by using such assigned metadata, it can be expected that the number of posted information that can automatically interpret the contents of the information and perform some kind of processing can be expected.

  FIG. 10 is a functional block diagram showing a schematic function of the metadata assignment support apparatus 600 according to the third embodiment. The metadata assignment support apparatus 600 includes a bulletin information acquisition unit 300, a dictionary holding unit 302, a word extraction unit 304, a lexical system holding unit 306, a candidate output unit 308, a metadata generation unit 310, and metadata. A holding unit 312, a condition holding unit 610, and an addition notification unit 612 are included.

  The bulletin information acquisition unit 300, the dictionary holding unit 302, the word extraction unit 304, the lexical system holding unit 306, the candidate output unit 308, the metadata generation unit 310, which have already been described as constituent elements in the first embodiment. The metadata holding unit 312 has substantially the same function, and thus a duplicate description thereof will be omitted. Here, the condition holding unit 610 and the addition notification unit 612 having different configurations will be mainly described.

  The condition holding unit 610 holds fields or metadata conditions desired by the user. The user registers specific metadata, for example, “event”-“venue”-“Tokyo exhibition hall” with the ID for identifying the user and the notification destination to the user, with respect to the metadata providing support device 600. To do. Such registration may be performed by enumerating metadata, or may be performed only by [Type]-[Attribute], [Type], or only by field, [Type], [Attribute], [Attribute Value]. ], For example, “skiing” and “snowboarding” may be combined and performed as “winter sports”. Furthermore, it is possible to use the logical sum or logical product of any of the field, [type], [attribute], and [attribute value] as a condition.

  Further, when the user assigns metadata to arbitrary posted information, the metadata assignment support apparatus 600 may register the automatically assigned metadata in the condition holding unit 610 as a metadata condition. . Usually, the user gives metadata to the posted information of interest. In the above configuration, the user can register the metadata of interest in the condition holding unit 610 at the same time as adding the metadata.

  When the metadata that satisfies the condition held in the condition holding unit 610 is added to the metadata holding unit by another user, the addition notification unit 612 notifies the registered user to that effect. For example, when the above metadata of “event”-“venue”-“Tokyo exhibition hall” is registered in the condition holding unit 610 as a condition for a specific user, other users post the same metadata. When it is added to the information, the fact is notified to the user who has registered the metadata, the fact that the posting information desired by the user has been registered, and the user is encouraged to view the posting information. Such notification may include a keyword and a URL (Uniform Resource Locator) indicating the posted information.

  A method for assigning metadata using the metadata assignment support apparatus 600 will be described below.

(Metadata grant support method)
FIG. 11 is a flowchart showing the processing flow of the metadata assignment support method according to the third embodiment. The processing flow will be described below together with the configuration of FIG.

  As a premise of the metadata providing support method, it is assumed that the condition holding unit 610 registers a field desired by the user or metadata conditions.

  First, the posting information acquisition unit 300 of the metadata assignment support apparatus 600 acquires the posting information 350 by the HTTP protocol via the communication network 200 (S700). Then, the word extraction unit 304 deletes a tag or the like unnecessary for specifying the document, decomposes the document included in the posting information 350 into a morpheme by morphological analysis or syntax analysis, and specifies a keyword for specifying the posting information 350; Words defined in the dictionary 352 that can represent [attribute values] of the keyword (for example, proper nouns such as personal names and place names) are extracted (S702).

  Subsequently, the candidate output unit 308 follows the association of the vocabulary system 354 with the keyword candidates extracted in the word extraction step (S702) and the word extracted by the word extraction unit 304, which should be attribute values of the keywords. Candidates are listed, and a Web document representing the candidate list in a menu format is created (S704). Then, the Web document is output (displayed) to the monitor of the terminal owned by the user or the monitor 356 of the metadata assignment support apparatus 230 (S706). The display format at this time is not limited to the menu format, and various selection formats can be applied.

  Next, the metadata generation unit 310 sets the word selected by the user as [attribute value], and makes a pair with the keyword and [attribute], that is, a pair of [type]-[attribute]-[attribute value]. The metadata 360 is generated (S708). Then, the metadata holding unit 312 stores the metadata 360 generated by the metadata generation unit 310 in a form that can be referred to from other terminal devices 220 (S710).

  Subsequently, the addition notification unit 612 compares the metadata generated by the metadata generation unit 310 with the metadata conditions registered by other users in the condition holding unit 610 (S712). This comparison is made according to the metadata conditions. If the metadata is simply listed, it is judged whether it is completely matched with the metadata. To be judged. As a result of the comparison, if the same metadata exists, the user who has registered the metadata is notified (S714).

  Also provided are a program for causing a computer to perform the above-described metadata assignment support method, and a storage medium storing the program.

  As described above, when a user designates (registers) a desired field or metadata so that the designated field or metadata is added to the posted information by another user, such posted information Can be obtained and / or the added bulletin information itself. Each user takes the initiative to receive information about the fields that they are interested in in order to receive information provided by other users on their own and newly published information, and to provide information to other users. As a result, the use of the grant support device is promoted, so that the posting information can be given metadata. By adding metadata to such a large amount of bulletin information, the utility of the entire system is increased, and a system that is easier to use for the user is constructed.

  As mentioned above, although preferred embodiment of this invention was described referring an accompanying drawing, it cannot be overemphasized that this invention is not limited to the example which concerns. It will be apparent to those skilled in the art that various changes and modifications can be made within the scope of the claims, and these are naturally within the technical scope of the present invention. Understood.

  For example, in the above-described embodiment, the creator of the posting information and the creator of the metadata are represented as different persons. However, the present invention is not limited to such a case. For example, the resource information server and the metadata providing support device are integrally formed. In addition, metadata can be added together with the posting information.

  In the above embodiment, the metadata holding unit for holding the metadata is provided in the metadata assignment support apparatus. However, the metadata holding unit may be provided separately, and a plurality of metadata DBs may be provided for regular use. It is also possible to configure to synchronize metadata.

  In the above embodiment, unnecessary tags are deleted in the word extraction unit. Conversely, it is also possible to extract words in a specific tag using a specific condition using this tag. it can. For example, it is possible to preferentially handle HTML or XML header tags as keywords.

  In the above embodiment, the description is limited to the posted information on the Internet in order to facilitate understanding. However, the present invention is not limited to this, and the Web service in which the corresponding document exists or other electronic information is described. It can also be applied to other documents. Further, even for an article that is not an intangible, for example, it can be applied to a system that searches for an article that satisfies a necessary condition with a service that uses a document and position information describing the article.

  Further, the target bulletin information is not limited to Japanese, and the present invention can deal with various languages such as alphabet and Korean.

  Furthermore, in the above embodiment, a plurality of vocabulary systems for each field are arranged in the same hierarchy, but fields can also be represented hierarchically. For example, a vocabulary system of “baseball” may be arranged in a hierarchy below the vocabulary system of the field “sports”. At this time, it is possible to limit the application range according to the situation.

  In the above embodiment, each configuration has been described by taking a search engine as an example. However, the concept of ontology and vocabulary can be applied to other applications, for example, a link such as an electronic calendar and a scheduler. Can also be used. For example, if there is a date etc. as metadata [attribute] for posted information such as events, the event information related to the relevant date is displayed in the calendar or scheduler, etc., so that it can be efficiently linked with external information. Thus, the information can be captured. In addition, by providing the application with an input / output mechanism that can interpret the metadata, various processes such as configuring an application that displays only the event information of the field of interest using the attributes of the theme and field. Easy to implement.

  Note that the steps in the metadata assignment support method of the present specification do not necessarily have to be processed in chronological order in the order described in the flowchart, but are performed in parallel or individually (for example, parallel processing or Object processing) may also be included.

It is explanatory drawing which showed an example of the vocabulary system. It is the text figure which showed an example of the document in bulletin information. It is explanatory drawing which showed the example of the metadata set based on the vocabulary system shown in FIG. It is a block diagram explaining the outline of an information collection system. It is a functional block diagram showing the schematic function of the metadata grant support device in a 1st embodiment. It is the flowchart which showed the flow of the process of the metadata provision assistance method in 1st Embodiment. It is the functional block diagram which showed the schematic function of the metadata provision assistance apparatus in 2nd Embodiment. It is the functional block diagram which showed the schematic function of the metadata provision assistance apparatus in the other example of 2nd Embodiment. It is the functional block diagram which showed the schematic function of the metadata provision assistance apparatus in the other example of 2nd Embodiment. It is the functional block diagram which showed the schematic function of the metadata provision assistance apparatus in 3rd Embodiment. It is the flowchart which showed the flow of the process of the metadata provision assistance method in 3rd Embodiment.

Explanation of symbols

230, 500, 600 Metadata assignment support device 304 Word extraction unit 306 Lexical system holding unit 308 Candidate output unit 310 Metadata generation unit 312 Metadata holding unit 354 Lexical system 510 Field selection unit 520 Desired field estimation unit 530 Posted field estimation unit 610 Condition holding unit 612 Addition notification unit

Claims (7)

Connected to the communication network,
A word extraction unit for extracting words from a document included in the posted information by natural language analysis;
A vocabulary holding unit that holds an ontology-based vocabulary;
A candidate output unit that outputs a keyword and a word candidate extracted by the word extraction unit, which should be an attribute value of the keyword, according to the association defined in the vocabulary system;
A metadata generation unit that generates metadata paired with a word selected by the user from the keywords, attributes, and candidates;
A metadata holding unit for holding the generated metadata;
An information collecting apparatus comprising: A field selection unit that allows the user to select a desired field;
The vocabulary system holding unit includes a plurality of vocabulary systems provided for each field,
The information collection device according to claim 1, wherein the candidate output unit outputs attribute value candidates according to a vocabulary system of a field selected by a user by the field selection unit. A desired field estimation unit for estimating a field desired by the user from an operation history of the user on the information collection device;
The vocabulary system holding unit includes a plurality of vocabulary systems provided for each field,
The information collection apparatus according to claim 1, wherein the candidate output unit outputs attribute value candidates according to a lexical system of the estimated field. A posting field estimation unit for estimating a field to which the posting information belongs based on the word extracted by the word extraction unit;
The vocabulary system holding unit includes a plurality of vocabulary systems provided for each field,
The information collection apparatus according to claim 1, wherein the candidate output unit outputs attribute value candidates according to a lexical system of the estimated field. A condition holding unit for holding metadata conditions desired by the user;
An additional notification unit for notifying the user when metadata satisfying the condition is added to the metadata holding unit by another user;
The information collection device according to claim 1, further comprising: A word extraction step of extracting words from the document included in the posted information by natural language analysis;
A candidate output step for outputting a keyword and a word candidate extracted by the word extraction step, which should be an attribute value of the keyword, in accordance with the association defined in the ontology-based vocabulary system;
A metadata generation step of generating metadata in which a word selected by the user from the keyword, attribute, and candidate is paired;
A metadata holding step for holding the generated metadata;
An information collection method characterized by including: Computer
A word extraction step of extracting words from the document included in the posted information by natural language analysis;
A candidate output step for outputting a keyword and a word candidate extracted by the word extraction step, which should be an attribute value of the keyword, in accordance with the association defined in the ontology-based vocabulary system;
A metadata generation step of generating metadata in which a word selected by the user from the keyword, attribute, and candidate is paired;
A metadata holding step for holding the generated metadata;
A program characterized by running

Images