JP4607830B2 - Interest information generating apparatus, interest information generating method, and interest information generating program - Google Patents

Description

  The present invention relates to an interest information generation apparatus, an interest information generation method, and an interest information generation program, and more particularly to a method of automatically generating a personal ontology in which personal interest information is conceptually hierarchized while referring to a blog entry. And suitable.

When performing an information search on the Internet or the like, an information search is mainly performed by inputting a keyword. In this keyword search, for example, in the field of video recording devices, since the recording content is specified by a keyword, the intended video content could not be recorded unless the user came up with an appropriate keyword.
Further, even in a search using a search engine such as google, if the user does not come up with an appropriate keyword, the intended content cannot be searched, and many unnecessary search results may be included.
Furthermore, even in blog providers such as doblog, the user remains at the level of searching for information of interest on a keyword basis and accessing the blog site according to the search result, and the unknown keyword or community in which the user is interested. Or I couldn't find a blog site.

One reason for this is that the current search method is only keyword search, and a high-precision search using the class (concept) system and class attributes, that is, class names and instances (substances) cannot be performed. It is done. The keyword is just a character string, but the ontology class has multiple instances as members, and the individual preference is reflected in the class depending on what instance is a member of the class. Can do. Also, the preference of each individual can be reflected in the class depending on the class hierarchy.
Here, in the conventional method for calculating the degree of approximation between ontology, as disclosed in Non-Patent Document 1, an ontology constructed in a completely distributed environment is targeted. For this reason, as a pre-process for calculating the degree of approximation between ontology, a correspondence relationship between classes of ontology is derived.

FIG. 6 is a diagram illustrating a configuration example of an ontology constructed in a conventional completely distributed environment.
In FIG. 6, the ontology OA has a class called “lock”. The class “Rock” has a child class “US”, the child class “US” has a child class “Indie”, and the child class “Indies” has a “Grunge” class. There is a child class. Further, it is assumed that instances “nirvana”, “Foo Fighters”, and “super chunk” exist in the child class “grunge”.

  On the other hand, the ontology OB has a class called “lock”. The class “Rock” has child classes “America” and “United Kingdom”, and the child class “America” has a child class “Indies” in the child class “America”. It is assumed that instances “nirvana” and “Stone temple pilots” exist in the child class “indies”.

Furthermore, the child class “UK” has a child class “Alternative”, and the child class “Alternative” has a child class “Scotland”. It is assumed that an instance of “tenage” exists in “Scotland”.
When the degree of approximation between the ontology OA and OB is measured, mapping between classes included in the ontology OA and OB is performed. Here, when mapping between classes included in the ontology OA and OB is performed, the degree of approximation can be measured for all combinations between classes constituting the ontology OA and OB.

  That is, as shown in FIG. 7 and FIG. 8, mapping between the ontology OA and OB can be performed by calculating the approximate degree between all classes included in each ontology OA and OB. . As a method of measuring the degree of approximation, it can be calculated based on the degree of approximation of class characteristics such as name attributes, various properties, and instance sets possessed by a class.

FIG. 9 is a diagram illustrating an example of a mapping result between the classes of FIG.
In FIG. 9, as a result of mapping between ontology OA and OB, ontology OA “lock” class and ontology OB “lock” class are mapped, ontology OA “US” class and ontology OB “USA” ”Is mapped, ontology OA“ indie ”class is mapped to ontology OB“ indies ”class, ontology OA“ grunge ”class and ontology OB“ indies ”class is mapped It shall be assumed.

  When mapping between the classes of ontology OA and OB is performed, the degree of approximation between the class sets composed of the adjacent classes of the class having a high degree of approximation is measured. As a method for measuring the degree of approximation between class sets, for example, a Jaccard coefficient can be used. With this Jaccard coefficient, the degree of approximation between the class sets X and Y can be obtained by (X∩Y) / (X∪Y).

FIG. 10 is a diagram illustrating an example of the approximation degree measurement result between the ontology in FIG.
In FIG. 10, it is assumed that the mapping result of FIG. 9 is obtained between the classes of ontology OA and OB of FIG. In this case, in ontology OA, a class set G11 consisting of a class called "lock" and a class called "US", a class called "lock", a class set G12 consisting of a class called "US" and a class called "indie", "indie" And a class set G13 including the class “grunge” is obtained.

  On the other hand, in ontology OB, a class set G21 consisting of a class "Rock", a class "America" and a class "UK", a class set "Class" "Rock", a class "America" and a class "Indies" G22, a class set G23 consisting of a class “America” and a class “indies” is obtained.

  Between the class sets G11 and G21, the class “lock” in the class set G11 and the class “lock” in the class set G21 are in a correspondence relationship. Since the class “US” and the class “USA” in the class set G21 are in a corresponding relationship, they are regarded as one member, and the class set G21 includes a class “UK”. The total number of members G11 and G21 (G11∪G21) is 3.

  On the other hand, the class “lock” in the class set G11 and the class “lock” in the class set G21 have a correspondence relationship, and the class “US” in the class set G11 and the class “America” in the class set G21 correspond to each other. Since there is no class corresponding to “UK” in the class set G21 in the class set G11, the number of members commonly included between the class sets G11 and G21 (G11∩G21) is two. As a result, the Jaccard coefficient between the class sets G11 and G21 is 2/3.

Similarly, the Jaccard coefficient between the class sets G12 and G22 is 2/4, and the Jaccard coefficient between the class sets G13 and G23 is 1/3.
When the Jaccard coefficients between the ontology OA and OB class sets are found, the Jaccard coefficients between these class sets are added together. Then, by dividing the result obtained by adding the Jaccard coefficients by the number of classes of the source ontology OA, the degree of approximation of the ontology OB viewed from the ontology OA can be measured. For example, in FIG. 10, since the number of classes of the source ontology OA is 4, the degree of approximation of the ontology OB viewed from the ontology OA is (2/3 + 2/4 + 1/3) / 4 = 3/8.

FIG. 11 is a diagram illustrating a comparative example of the degree-of-approximation measurement results for a plurality of conventional ontology.
In FIG. 11, it is assumed that the ontology OC has no child class below the class “UK” of the ontology OB. In the ontology OC, a class set G31 composed of a class “Rock” and a class “America”, a class set G32 composed of a class “Lock”, a class “America” and a class “indies”, “America” A class set G33 including the class and the class “indies” is obtained.

In this case, the Jaccard coefficient between the class sets G11 and G31 is 2/2, the Jaccard coefficient between the class sets G12 and G32 is 2/4, and the Jaccard coefficient between the class sets G13 and G33 is 1/3. As a result, the degree of approximation of ontology OC viewed from ontology OA is (2/2 + 2/4 + 1/3) / 4 = 11/24.
Non-Patent Document 2 discloses a method for realizing semi-automatic mapping between classes constituting a plurality of ontology.

Maedche, A .; and Staab, S .; : Measuring Similarity between Ontologies, In Technical Report, E0448, University of Karlsruhe (2001) Makoto Nakajo, Yu Miyoshi, Yasuyuki Kimura: “Proposal and Evaluation of Message Mapping Method Based on Semantic Information for Flexible System Cooperation”, Database Society of Japan Letters, Vol. 4, no. 1, pp. 37-40 (2005)

  However, since it is expensive to generate a personal ontology in which the user's interests are conceptually hierarchized, it is difficult to distribute each person's personal ontology via the Internet. For this reason, by mapping the personal ontology of each individual to the personal ontology of another user, it has not been possible to automatically form a community between users with whom the interests match. In addition, there is a problem that automatic and highly accurate information retrieval and recommendation cannot be performed by matching between metadata of content such as music files on the Web and personal ontology.

In addition, in the conventional personal ontology generation method, users who are eagerly talking about an artist may be, but one user is eagerly talking about the artist, and another user is not eagerly talking about the artist. In some cases, since the degree of interest in the artist is measured to be the same, there is a problem that a personal ontology that sufficiently reflects the user's interest cannot be generated with high accuracy.
Accordingly, an object of the present invention is to provide an interest information generation apparatus, an interest information generation method, and an interest information generation program capable of generating a personal ontology in which an individual's interest is reflected with high accuracy while suppressing the cost of creation. Is to provide.

から計測し、前記一のブログエントリに含まれる単語を含むクラスに対する前記興味度を、

から計測することを特徴とする。 To solve the problems described above, according to the interest information generating apparatus according to claim 1, storage means for storing a blog entry for each individual, a word extraction means for extracting words included in the blog entry, previously Storage means for storing a template ontology in which set words are conceptually hierarchized; classifier applying means for extracting a class or instance including a word extracted by the word extracting means from the template ontology; and the blog entry Based on the degree of interest measured by the interest degree measuring means, the interest degree measuring means for measuring the degree of interest of the individual having the blog entry, with respect to the class ontology class or instance including the word included in the degree of interest for each individual for a relatively large the class or instance, the class or Instruments Includes Manual and a hierarchical structure that includes a class of their upper, and personal ontology extracting means for extracting from the model ontology as a personal ontology representing the interest information of the individual, and the interest degree measuring means, each individual of the blog entry The instance including the word included in the one blog entry, where E is the set of and the class and instance type including the word appearing in one blog entry of the set E is N (Ei) The degree of interest in

And the degree of interest in the class including the word included in the one blog entry,

It is characterized by measuring from .

から計測し、前記一のブログエントリに含まれる単語を含む前記クラスに対する興味度を

から計測するステップと、計測された前記興味度に基づいて、前記興味度が比較的大きな前記クラスまたは前記インスタンスについて、当該クラスまたはインスタンスおよびそれらの上位のクラスを含む階層構造を、個人の興味情報を表すパーソナルオントロジとして前記雛形オントロジから個人ごとに抽出するステップと、を備えることを特徴とする。 In addition, according to the interest information generation method according to claim 2, the interest information generation method executed by the interest information generation apparatus that generates, as interest information, a personal ontology in which words included in a personal blog entry are hierarchized into a concept hierarchy. A step of extracting words included in the blog entry stored for each individual in the storage means, and from a template ontology stored in the storage means and preliminarily hierarchized in terms of concepts, from the blog entry Extracting a class or instance including the extracted word, a set of individuals of the blog entry E, and the number of types of the class and instance including a word appearing in one blog entry of the set E When N (Ei), the individual having the one blog entry is included in the one blog entry That the degree of interest with respect to the instance that contains the word,

And measure the degree of interest in the class containing the words included in the one blog entry.

A step of measuring the, on the basis of the degree of interest that is measured, for the degree of interest is relatively large the class or the instances, the hierarchical structure including the class or instance and class of their upper, personal interest information , extracting for each individual from the template ontology as a personal ontology representing the characterized in that it comprises a.

から計測し、前記一のブログエントリに含まれる単語を含む前記クラスに対する興味度を、

から計測するステップと、計測された前記興味度に基づいて、当該興味度が比較的大きな前記クラスまたは前記インスタンスについて、当該クラスまたはインスタンスおよびそれらの上位のクラスを含む階層構造を、個人の興味情報を表すパーソナルオントロジとして前記雛形オントロジから個人ごとに抽出するステップと、をコンピュータに実行させることを特徴とする。 Further, according to the interest information generating program according to claim 3, the step of extracting words included in the blog entry stored for each individual in the storage means, and the preset words stored in the storage means Extracting a class or an instance including a word extracted from the blog entry from the model ontology formed into a form, and an individual set of the blog entry to appear in one blog entry of the set E When the number of types of classes and instances that include words to be N (Ei) is, the degree of interest of the individual having the one blog entry with respect to the instance that includes the words included in the one blog entry,

The degree of interest in the class including the words included in the one blog entry,

And, based on the measured degree of interest, for the class or the instance having a relatively high degree of interest, a hierarchical structure including the class or the instance and higher classes thereof is used as the personal interest information. And extracting the individual ontology from the template ontology for each individual as a personal ontology representing

  As described above, according to the present invention, it is possible to search for information suitable for one's preference based on the consistency between concepts rather than just a character string, and consider the degree of personal interest. However, it is possible to generate a personal ontology that accurately reflects personal interests while collating the words included in each individual's interest information on the template ontology, while eliminating the misclassification of ambiguous words. it can. For this reason, it is possible to generate personal ontology with high accuracy while suppressing the cost of creation, and it is possible to improve the accuracy of information retrieval, and to distribute each person's personal ontology widely on the Internet. And it is possible to form a community that suits individual preferences.

Hereinafter, an interest information generation apparatus and method according to an embodiment of the present invention will be described with reference to the drawings.
FIG. 1 is a block diagram showing a schematic configuration of a system to which an interest information generating apparatus according to an embodiment of the present invention is applied.
In FIG. 1, terminals 2 to 4 and a server 5 are connected via a communication network 1. As the communication network 1, for example, a public communication network that performs IP communication can be used, and the Internet may be used. Further, it may be a private communication network between companies or a public communication network, but it is like an IP-VPN (Internet Protocol-Virtual Private Network) that can provide dedicated communication with high reliability and security. A simple net may be used. Further, the terminals 2 to 4 may be a notebook personal computer or a desktop personal computer, and may be a mobile phone terminal or a PDA (Personal Data Assistant). The server 5 can be installed on a blog provider or ISP (Information Service Provider). As the server 5, for example, a ping server that collects and provides update information of a blog can be used.

  Here, the server 5 is provided with blog sites 7 to 9 corresponding to the terminals 2 to 4 respectively, and the blog entries 7a to 7n, 8a to 8n, and 9a to 9n are held in the blog sites 7 to 9, respectively. Has been. The blog entries 7a to 7n, 8a to 8n, and 9a to 9n represent minimum units of articles in the blog, and can be provided for each day. The server 5 holds a template ontology 6, and the template ontology 6 is provided with a personal ontology template in which personal interest information is conceptually hierarchized.

  The template ontology 6 can be arbitrarily created on the blog provider side. For example, if the blog provider wants the users of the terminals 2 to 4 to construct a personal ontology related to music, the template ontology 6 related to music may be constructed. Here, in order to express the interests of the users of the terminals 2 to 4 in detail, it is preferable to construct a template ontology 6 that is subdivided as much as possible and has high coverage. The entity of the template ontology 6 is a text file described in an XML language such as the ontology description language OWL. In order to simplify the organization of information, instances may be classified only into the lowest class.

  Furthermore, the server 5 applies words that appear frequently in the blog entries 7a to 7n, 8a to 8n, and 9a to 9n by applying morphological analysis to the blog entries 7a to 7n, 8a to 8n, and 9a to 9n, respectively. In the frequent word extraction means 5a to be extracted, the classifier application means 5b and the classifier application means 5b to extract from the template ontology 6 a class or instance including words that frequently appear in the blog entries 7a to 7n, 8a to 8n, and 9a to 9n. Interest level measuring means 5c and interest level measuring means 5c for measuring the degree of interest in these classes or instances based on the ratio of the extracted classes or instances included in the blog entries 7a to 7n, 8a to 8n, 9a to 9n. Class or instance and all classes above it based on the interest measured in Personal ontology extraction means 5d for extracting from template ontology 6 is provided as a personal ontology.

  The frequent word extraction means 5a applies morphological analysis to the blog entries 7a to 7n, 8a to 8n, and 9a to 9n, respectively. Then, morphemes that frequently appear in a plurality of blog entries 7a to 7n, 8a to 8n, and 9a to 9n possessed by the same user are extracted. At this time, morphemes that are obviously not of interest (e.g., me, things, etc.) can be filtered.

  Next, the classifier applying unit 5b applies each morpheme frequently appearing in the blog entries 7a to 7n, 8a to 8n, and 9a to 9n to the template ontology 6, and a character string that matches the class or instance in the template ontology 6 is applied. Find out if there is. When the character string matching the class or instance in the template ontology 6 is in the blog entries 7a to 7n, 8a to 8n, and 9a to 9n, the interest degree measuring means 5c has the class or instance of the blog entry 7a to Based on the ratio included in 7n, 8a to 8n, and 9a to 9n, the degree of interest in those classes or instances is measured. Then, the personal ontology extraction means 5d obtains the direct descendant classes and instances from the root class of the template ontology 6 to the class or instance with respect to the class or instance having a high degree of interest measured by the interest degree measurement means 5c. Extract as a personal ontology.

  This makes it possible to search for information that suits one's preference based on the consistency between concepts rather than just character strings, and is included in each individual's interest information while taking into account the degree of personal interest By collating words on the template ontology 6, it is possible to generate a personal ontology in which personal interests are accurately reflected while eliminating the misclassification of ambiguous words. For this reason, it is possible to generate personal ontology with high accuracy while suppressing the cost of creation, and it is possible to improve the accuracy of information retrieval, and to distribute each person's personal ontology widely on the Internet. And it is possible to form a community that suits individual preferences.

Here, the degree of interest in a class or instance can be defined as follows, for example.
(1) The degree of interest of the user per entry is 1.
(2) If the types of user classes and instances appearing in an entry E _i are N (E _i ), the degree of interest of each class and instance of the user in the entry E _i is 1 / N (E _i ). is there.
(3) The degree of interest for each instance I _i in the ontology is given by the following equation, where E is the total accumulated entry set of the user.

(4) The degree of interest for each class C _i in the ontology is given by the following equation, where E is the total accumulated entry set of the user.

(5) The interest level of the instance is inherited by the interest level of the class to which the instance belongs, and the interest level of the child class is inherited by the interest level of the parent class.
FIG. 2 is a diagram illustrating an example of the degree of interest of classes and instances according to an embodiment of the present invention.
In FIG. 2, it is assumed that entries E ₁ and E ₂ are included in all the stored entries of user A. In the interest ontology OA of the user A, a class called “alternative” exists, and a class called “alternative” has child classes called “Madchester” and “Shoegaze”. The class “alternative” has an instance “nirvana”, the class “Madchester” has instances “New Order” and “Stone Roses”, and the class “Shoegaze” has “ Assume that instances of “My Bloody Valentine” and “Ride” exist.

In entry E ₁ , four artists “Stone Roses”, “My Bloody Valentine”, “nirvana”, and “New Order” appear, and in entry E ₂ , one artist “My Bloody Valentine” appears again. Above, one genre “Shoegaze” and one artist “Ride” appear.
Therefore, the interest level of the instance “New Order” is 1/4, the interest level of the instance “Stone Roses” is 1/4, and the interest level of the instance “My Bloody Valentine” is 1/4 + 1/3 = 7 / 12. The interest level of the instance “Ride” is 1/3, and the interest level of the instance “nirvana” is 1/4.

  Further, the interest level of the class “Madchester” can be obtained by adding the interest levels of the instances “New Order” and “Stone Roses” under the class, and 1/4 + 1/4 = 1/2. It becomes. In addition, the interest level of the class “Shoegaze” can be obtained by adding the interest levels of the instances “My Bloody Valentine” and “Ride” under the class, and 7/12 + 1/3 = 5/4 It becomes. Further, the interest level of the class “alternative” can be obtained by adding the interest levels of the instances “nirvana”, “Madchester” and “Shoegaze” subordinate to the class “1/2 + 5/4 + 1”. / 4 = 2.

As a result, it is possible to recognize that the user A's interest is strong because the artist “My Bloody Valentine” has been written in the two entries E ₁ and E ₂ . In addition, regarding the genre of “Shoegaze”, the class name also appears in entry E ₂ and the subordinate artists are frequently spoken in entries E ₁ and E ₂ , so the degree of interest can be increased. Even if there are many artists under their control, it can be prevented that they are treated equally as interests.

  Note that the template ontology 6 may be manually created by desk work and stored in the server 5, or the personal ontology extracted by the personal ontology extracting means 5c may be merged with the existing template ontology 6. You may create it. Furthermore, a class or instance that the user is interested in may be added to the personal ontology extracted from the template ontology 6, or a class or instance that the user is not interested in may be deleted.

The frequent word extraction means 5a, classifier application means 5b, interest degree measurement means 5c, and personal ontology extraction means 5d cause the computer to execute a program in which instructions for performing the processing performed by these means are described. Can be realized.
If this program is stored in a storage medium such as a CD-ROM, the frequent word extraction means 5a and classifier application means are installed by installing the storage medium in the computer of the server 5 and installing the program in the computer. 5b, the processing performed by the interest degree measuring means 5c and the personal ontology extracting means 5d can be realized. Moreover, by downloading this program via the communication network 1, this program can be easily spread.

  In addition, when a computer executes a program in which an instruction for performing processing performed by the frequent word extraction unit 5a, the classifier application unit 5b, the interest degree measurement unit 5c, and the personal ontology extraction unit 5d is executed by the stand-alone computer Alternatively, it may be distributed to a plurality of computers connected to the network.

FIG. 3 is a diagram illustrating a method for generating a personal ontology according to an embodiment of the present invention.
In FIG. 3, an entry set of users A, B,..., X is collected through a ping server or the like, and an index is created by performing morphological analysis on all the collected blog entries ( Step S1).
Next, all the blog entries collected by the ping server are classified with respect to the template ontology OH (step S2). Here, as a method for classifying blog entries, if there is a description in an entry having a name attribute of class C _{i having} a template ontology OH, the entry can be classified into class C _i . Also, if the name attribute of the instance I _i (∈C _i ) belonging to the class C _i having the template ontology OH is in the description in the entry, the entry is classified into the instance I _i belonging to the class C _i. Can do. Note that the same entry may be classified into a plurality of classes.

For example, when a character string “Charlantans” is included in the description in the entry, the entry can be classified into an instance “Charlantans” of the class “Madchester”.
Next, the number of users A, B,..., X who are interested in each instance of the lowermost class C ₁ forming the template ontology OH is measured (step S3). When the number of users A, B,..., X who are interested in instances of class C _l is measured, the number of users even when the same user describes instance I _l in a plurality of entries. Measures 1

Next, the user A interested in lowermost class C _l to form a stationery ontology OH, measured B, · · ·, the number of X. Here, the number of users A, B,..., X interested in the lowest class C ₁ forming the template ontology OH is the number of users interested in all instances under the lowest class C _1. , it can be calculated by the sum of the number of users interested in the lowest layer class C _l itself. Note that the number of users is measured as 1 even when the same user is interested in a plurality of instances or is interested in the lowermost class and instances belonging to the lowermost class at the same time. In this way, by recursively measuring the number of users A, B,..., X interested in the class or instance forming the template ontology OH up to the root class, the user A interested in the domain , B,..., X can be calculated.

Next, when all the blog entries collected by the ping server are classified with respect to the template ontology OH, the classification results are arranged for each user ID, whereby each user A, B,. Each interest ontology OA,..., OX is generated (step S4).
Here, when all the blog entries collected by the ping server are classified with respect to the template ontology OH, the instances that belong to the same class of the ontology have the same property, and between classes close to the class hierarchy The classification error can be eliminated by using the property that the properties of are close and the properties of both instances are also close.

FIG. 4 is a diagram illustrating a method for measuring the degree of approximation between personal ontology according to an embodiment of the present invention. In the following description, when the degree of approximation between an ontology and another ontology is measured, the former is called a source ontology and the latter is called a target ontology.
In FIG. 4, classes “b1”, “b2”, and “b3” exist immediately below the class “a1” of the template ontology OH, and “c1” and “c3” immediately below the class “b1”. It is assumed that a class called “c2” exists and classes “d1” and “d2” exist immediately below the class called “c1”. Also, it is assumed that classes “c3” and “c4” exist immediately below the class “b2”, and a class “c5” exists immediately below the class “b3”.

  The class “d1” includes instances “j” and “k”, the class “d2” includes the instance “l”, and the class “c2” includes “m”. "B", the class "b2" has an instance "n", and the class "c3" has "a", "e", "c", "f", "b" ”,“ D ”, and“ g ”exist, and the class“ c4 ”includes instances“ p ”,“ g ”,“ j ”, and“ h ”.

Then, the words that frequently appear in each user's blog entry are extracted, and the classes or instances including the words and all the classes above them are extracted from the template ontology OH, thereby creating personal ontologies OA and OB. Shall be.
Here, the classes “b1” and “b2” exist immediately below the class “a1” of the personal ontology OA, and the classes “c1” and “c2” exist immediately below the class “b1”. It is assumed that a class “d1” exists immediately below the class “c1”, and a class “c3” and “c4” exist immediately below the class “b2”. , The class “d1” includes instances “j” and “k”, the class “c2” includes the instance “m”, and the class “c3” includes “a”. , “C”, “b”, and “d” exist, and the class “c4” includes instances “q” and “h”.

  Also, in the personal ontology OB, there are “b1”, “b2”, and “b3” classes directly under the “a1” class, and there is a “c1” class directly under the “b1” class. Assume that a class “d1” exists immediately below the class “c1”, and a class “c3” and “c4” exist immediately below the class “b2”. The class “d2” has an instance “l”, the class “b2” has an instance “n”, and the class “c3” has “a” and “c”. ”,“ E ”, and“ f ”exist, and“ c4 ”class includes“ p ”and“ j ”.

  Then, a common class excluding a terminal class is analyzed between the template ontology OH and the personal ontology OA and OB, and a topology composed of parent and child classes with the common class as a parent class is extracted. In the example of FIG. 3, the end classes are “d1”, “d2”, “c3”, and “c4”. As a result, a child class set G1 having the class “a1” as the parent class, a child class set G2 having the class “b1” as the parent class, and the class “c1” being the parent class between the template ontology OH and the personal ontologies OA and OB. Child class set G3, child class set G4 having class “b2” as a parent class, and child class set G5 having class “b3” as a parent class can be extracted.

  The analysis of the common class only needs to confirm whether the same class ID exists between the template ontology OH and the personal ontology OA, OB. For this reason, it is not necessary to measure the degree of approximation of class name attributes, instance set properties, etc., and the amount of calculation can be reduced while maintaining the correspondence between classes accurately.

Next, the degree of approximation between the child class sets X and Y forming each topology between the personal ontology OA and OB is calculated with depth priority. Here, if the set of child classes of the classes constituting the template ontology OH is Z, the degree of approximation between the child class sets X and Y can be obtained by | X∩Y | / | Z |. By adding the approximation degree between the child class set for each topology, it is possible to measure the degree of approximation S _T personal ontology OA, between OB topology.

  For example, in the child class set G1, since the template ontology OH includes child classes “b1”, “b2”, and “b3”, the number of members of the child class set of the template ontology OH in the child class set G1 is 3 In the child class set G1, the personal ontology OA includes child classes “b1” and “b2”, and the personal ontology OB includes child classes “b1”, “b2”, and “b3”. Therefore, the child classes commonly included in the personal ontologies OA and OB are only “b1” and b2 ”, and the number of members of the product set of the child class sets of the personal ontology OA and OB in the child class set G1 is 2. As a result, the degree of approximation between the child ontology sets G1 of the personal ontology OA and OB is 2/3.

Similarly, the degree of approximation between the personal ontology OA and OB between the child class sets G2 is 1/2, the degree of approximation between the child class sets G3 is 0/2, and the degree of approximation between the child class sets G4 is 2/2. As a result, personal ontology OA, similarity _{S T} topology between OB becomes 2/3 + 1/2 + 0/2 + 2/2.
In the calculation of the approximate degree of topology between the personal ontology OA and OB, the class ID assigned from the template ontology OH is referred to, and the one along the connection form of the template ontology OH exists between the personal ontology OA and OB. You only have to confirm whether you want to do it. For example, whether the connection form a1-b1-c1 in the personal ontology OA, OB is the same as the connection form of the template ontology OH depends on whether the classes “a1”, “b1”, and “c1” are the personal ontology OA, OB. It can be determined simply by confirming whether or not it is held at.

  For this reason, in order to confirm the degree of coincidence of the topology between the personal ontology OA and OB, it is only necessary to check the class ID possessed by the personal ontology OA and OB. Therefore, it is possible to reduce the calculation amount of the degree of approximation of the topology between the personal ontology OA and OB.

Next, the degree of approximation between the common classes of the personal ontology OA and OB is calculated. Here, when calculating the degree of approximation between the common classes of the personal ontology OA and OB, an instance set belonging to the class can be used. That is, in a certain class C1, if the instance set of the source ontology is x, the instance set of the target ontology is y, and the instance set of the template ontology OH is z, the degree of approximation between the common classes of the personal ontology OA and OB is | x ∩y | / | z |. By adding the degree of approximation between the common class, it is possible to measure the degree of approximation S _C between personal ontology OA, the OB classes.

  For example, in the common class “b2” between the personal ontologies OA and OB, the instance set G6 in the common class “b2” of the personal ontology OA includes the instance “n”, and the common class “b2” of the personal ontology OB. An instance “n” exists in the instance set G6, and an instance “n” exists in the common class “b2” of the template ontology OH. As a result, the number of members of the common class “b2” of the template ontology OH is 1, the number of members of the product set of the instance set G6 of the personal ontology OA and OB is 1, and the degree of approximation between the common classes “b2” is 1/1.

In the common class “c2” between the personal ontology OA and OB, the instance set G7 in the common class “c2” of the personal ontology OA includes instances “a”, “c”, “b”, and “d”. In the instance set G7 in the common class “c3” of the personal ontology OB, instances “a”, “c”, “e”, and “f” exist, and in the common class “c3” of the template ontology OH “A”, “e”, “c”, “f”, “b”, “d”, and “g” exist.
As a result, the number of members of the common class “c3” of the template ontology OH is 7, the number of members of the product set of the instance set G7 of the personal ontologies OA and OB is 2, and the degree of approximation between the common classes “c3” is 2/7.
Therefore, the degree of approximation _{S C} between personal ontology OA, the OB class is 1/1 + 2/7.

Note that even in the calculation of the degree of approximation between the common classes of the personal ontology OA and OB, it is only necessary to confirm whether the instance ID allocated from the template ontology OH exists between the personal ontology OA and OB. For this reason, in order to calculate the degree of approximation between the classes of the personal ontology OA and OB, it is not necessary to confirm in advance the correspondence between instances based on the consistency of the names of instances, and the amount of calculation can be reduced.
Then, the personal ontology OA, the similarity S _C between the degree of approximation S _T and class topology in OB is determined, by using an evaluation function f (X) corresponding to the importance for the topology and class, the following expression The degree of approximation S _O (AB) between the personal ontology OA and OB can be given.
S _O (AB) = S _T + f (S _C )

  This makes it possible to construct personal ontology OA and OB while inheriting the class characteristics possessed by the template ontology OH, and directly measure the co-occurrence between instance sets to the degree of approximation between the personal ontology OA and OB. It is possible to appropriately extract personal ontologies OA and OB that suit individual tastes while reducing the amount of calculation, and at the same time, the domain ontology OH has knowledge of the domain that the personal ontology OA and OB have. It can be used as a relative measure of possession, and users who have a lot of knowledge about the domain can be effectively narrowed down.

The personal ontology OA, when measuring the degree of approximation S _T topology in OB, based on the number of members of the class set stationery ontology, personal ontology OA, by measuring the degree of approximation topology between OB, personal ontology Even when the number of members of the class set of OA and OB increases, the degree of topology approximation between the personal ontology OA and OB can be prevented from decreasing, and the amount of knowledge is large in the sense that there are abundant classes. The degree of approximation with the personal ontology OA, OB can be increased.

Here, it is possible to improve the accuracy of the degree of approximation between the personal ontology OA and OB by generating the personal ontology while considering the user's degree of interest, and efficiently acquire information close to the user's interest. be able to.
In other words, by introducing the user's degree of interest, if the user simply lists artists in the same blog entry, it can be considered that the interest in the artist appearing in the blog entry is low In addition, when the user describes an artist across different blog entries, it is possible to grasp that the interest in the artist appearing in the blog entry is high.

  For this reason, when detecting the user B who has an interest close to that of the user A based on the degree of approximation between the personal ontology OA and OB, the artist is simply enumerated in the blog entry, or the artist spans multiple blog entries. For user A who is only talking to one blog entry while describing the user A, it is possible to limit the user A's interest to a narrow range, and user A's personal ontology OA, user B's personal ontology OB, It is possible to reduce the degree of approximation between and reduce the recommended accuracy of interest information.

For example, in FIG. 2, the entry _{E 1, "Stone Roses",} "My Bloody Valentine", "nirvana" and "New Order" 4 artist appeared that, "My Bloody Valentine" artist that again an entry _{E 2} If it appears, the artist “My Bloody Valentine” can be treated as being highly interested by user A, and four artists “Stone Roses”, “My Bloody Valentine”, “nirvana”, and “New Order” are displayed. Can be prevented from being treated equally as the interest of the user A.
For this reason, it is possible to improve the accuracy of generating the interest ontology of the user A, improve the accuracy of measuring the degree of approximation between the interest ontologies, and improve the recommended accuracy of interest information.

  In FIG. 5, it is assumed that the blog entries PA and PB of the users A and B are classified with respect to the template ontology, thereby generating the interest ontologies KA and KB of the users A and B, respectively (step S11). . Then, the degree of approximation between the interest ontologies KA and KB of the users A and B is measured (step S12), and the topologies are analyzed by analyzing classes and instances that co-occur between the interest ontologies KA and KB having a high degree of approximation. Although it is different, information that is highly likely to be of interest can be recommended to the user as unexpected information through the entry of another user (step S13).

For example, by measuring the degree of approximation between the interest ontologies KA and KB of each user A and B, a user who is interested in a class such as “Madchester” or an instance such as “Happy Mondays” can have a class “Glasgow” It can be seen that there is a high possibility of being interested in an instance of “Teenage Fanclub”.
Moreover, by applying such interest ontologies KA and KB to a blog, it is possible to support the formation of a community based on an unexpected entry recommendation based on the degree of approximation between the interest ontologies KA and KB rather than a simple keyword search. The user's interest can be naturally expanded (step S14).

  In addition, as a method of measuring the degree of approximation between the interest ontology KA and KB, inference learning is performed based on the degree of approximation between class attributes such as class names and instances possessed by classes or the degree of topology approximation that is a connection form between classes. You can collate and measure the semantic approximation between classes of ontology. Alternatively, inference learning may be collated based on both the degree of approximation between class attributes and the degree of topology, and the semantic degree of approximation between classes of the ontology may be measured.

  Further, in the embodiment described above, a template ontology is necessary to automatically generate an interest ontology KA, KB. In order to design the template ontology, it is necessary to adjust the granularity of the end classes to reflect the hierarchical relationship between classes and the user's interest in detail. Here, topic directories in portal sites such as goo are becoming more detailed. For example, in the genre of the music domain, the granularity information of the genre hierarchy information published on the Web is finely set in consideration of the search according to the user's interest. For this reason, a template ontology can be constructed by using a topic directory in a portal site on the Internet.

  The present invention makes it possible to easily create a personal ontology and quickly obtain information that exactly matches one's interest, and automatically and efficiently obtain information that matches one's interest from an information source of an information communication system. Can be used.

1 is a block diagram illustrating a schematic configuration of a system to which an interest information generation apparatus according to an embodiment of the present invention is applied. It is a figure which shows an example of the interest degree of the class and instance which concern on one Embodiment of this invention. It is a figure which shows the production | generation method of the personal ontology based on one Embodiment of this invention. It is a figure which shows the approximation measuring method between personal ontology concerning one Embodiment of this invention. It is a figure which shows the community formation method using the approximation of the interest ontology which concerns on one Embodiment of this invention. It is a figure which shows the structural example of the ontology constructed | assembled in the conventional completely distributed environment. It is a figure which shows an example of the approximation measuring method by the brute force method between the classes of FIG. It is a figure which shows an example of the approximation measuring method by the brute force method between the classes of FIG. It is a figure which shows an example of the mapping result between the classes of FIG. It is a figure which shows an example of the approximation degree measurement result between the ontology of FIG. It is a figure which shows the comparative example of the approximation degree measurement result with respect to the conventional some ontology.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Communication network 2-4 Terminal 5 Server 5a Frequent word extraction means 5b Classifier application means 5c Interest degree measurement means 5d Personal ontology extraction means 6 Model ontology 7-9 Blog site 7a-7n, 8a-8n, 9a-9n Blog entry

Claims (3)

Storage means for storing blog entries for each person;
Word extraction means for extracting words contained in the blog entry ;
Storage means for storing a template ontology in which preset words are hierarchized in concept;
A classifier applying means for extracting a class or instance including words extracted by the word extraction means from the template ontology,
An interest degree measuring means for measuring an interest degree of an individual having the blog entry for the class or instance of the template ontology including the word included in the blog entry ;
Based on the degree of interest measured by the degree-of-interest measurement means, for the class or instance having a relatively large degree of interest , a hierarchical structure including the class or instance and higher classes thereof is obtained as personal interest information. Personal ontology extraction means for extracting from the template ontology for each individual as a personal ontology representing ,
The degree-of-interest measurement means, when the set of individuals of the blog entry is E, and the number of types of classes and instances including words that appear in one blog entry of the set E is N (Ei), The degree of interest with respect to the instance including the word included in the one blog entry,

And the degree of interest in the class including the word included in the one blog entry,

An interest information generation device characterized by measuring from An interest information generation method executed by an interest information generation apparatus that generates, as interest information, a personal ontology in which words included in a personal blog entry are conceptually hierarchized,
Extracting words contained in the blog entry stored for each individual in the storage means ;
Extracting a class or instance including a word extracted from the blog entry from a template ontology that is stored in a storage means and is configured by conceptually stratified words;
When the set of individuals of the blog entry is E and the number of types of classes and instances including words that appear in one blog entry of the set E is N (Ei), the blog entry has the one blog entry. The degree of interest of an individual in the instance containing the word contained in the one blog entry is:

And measure the degree of interest in the class containing the words included in the one blog entry.

Step to measure from ,
Based on the measured degree of interest, for the class or instance having a relatively high degree of interest , a hierarchical structure including the class or instance and higher classes is used as a personal ontology representing personal interest information. interest information generating method characterized by comprising the steps of: extracting for each individual from template ontology. Extracting words contained in the blog entry stored for each individual in the storage means ;
Extracting a class or instance including a word extracted from the blog entry from a template ontology that is stored in a storage means and is configured by conceptually stratified words;
When the set of individuals of the blog entry is E and the number of types of classes and instances including words that appear in one blog entry of the set E is N (Ei), the blog entry has the one blog entry. The degree of interest of an individual in the instance containing the word contained in the one blog entry is:

The degree of interest in the class including the words included in the one blog entry,

Step to measure from ,
Based on the measured degree of interest, for the class or instance having a relatively high degree of interest, the hierarchical structure including the class or instance and higher classes thereof is used as a personal ontology representing personal interest information. interest information generation program for causing and a step of extracting the personal each from model ontology, to the computer.

Images