KR20090078986A - Ontology-based semantic annotation system and method thereof - Google Patents

Abstract

A semantic annotation system based on ontology and a method thereof are provided to subdivide document information into the component unit of a document, not a document unit, and provide annotation with a semantic information knowledge base composed of ontology providing semantic mutual operability to the document, thereby making document search based on meaning possible. A semantic information administration unit(31) stores semantic information in a ontology storage to build a semantic information knowledge base. A document management unit(32) calls a document and attaches the semantic information to the document. If the semantic information related to the components of the document exists in the semantic information knowledge base, an auto annotation processor(33) annotates to the components of the document with the semantic information. An annotation preparation unit(34) annotates the semantic information to the components of the document for the components of document and semantic information which a user appoints.

Description

Ontology-based Semantic Annotation System and Its Method {Ontology-based Semantic Annotation System and Method}

The present invention belongs to the technical field of a semantic annotation system and method for constructing a semantic information knowledge base structured by constructing semantic information into an ontology and annotating a document using the semantic information knowledge base.

In addition, the present invention relates to a semantic annotation system and method for subdividing document information into document components instead of document units and using the semantic knowledge base to find and automatically annotate the document components suitable for the annotation. Belongs to the technical field.

In general, web technology expresses information in a simple markup language so that mutual information can be shared in the Internet space, thereby creating a breakthrough in network-based information sharing and delivery, and enormous information that interconnects various information resources. Spaces have been built, and as a result, the Internet has spread rapidly into real life, revolutionizing the information society. However, since the conventional web technology can search necessary information only by keyword, unnecessary information is also searched by simple keyword search. This is due to a structural problem in which the conventional web technology is a human-centered information processing technology, and the computer does not provide enough functions to effectively extract, interpret, and process information necessary for itself. In particular, as the amount of web information increases, this problem becomes more noticeable.

To solve this problem, the semantic web is a technology that can complement the existing web technologies to ensure semantic interoperability so that computers can understand them and to build an effective cooperative system between humans and computers. ) Came up. The semantic web provides a structure that enables not only humans but also machines to understand the information on the web, and supports semantic-based searching that finds only the results that are appropriate for a user's search request. It is a web technology that is proposed to enable automatic service on behalf of people by smoothly collaborating with.

In other words, the Semantic Web is a next-generation web technology that enables computers to understand, automate, integrate, and reuse the meaning of information resources. Ontology, semantically annotated web, and agent It consists of three main elements. Ontology is a formal specification system for shared conceptualization and provides semantic information of domain vocabulary. Ontology is a kind of knowledge expression, and the computer can understand the concept represented by the ontology and process the knowledge. In order to deal with inferences, the ontology's axiom and rule system are needed. A semantically annotated web is a knowledge base. The Semantic Web can build a huge knowledge base that semantically integrates the distributed information resources of the Internet. In a narrow sense, it may be possible to build a knowledge base of information resources of a company or institution. An agent is an intelligent agent that collects, retrieves and infers information resources on behalf of humans (users), and exchanges information with other agents. Intelligent agents are the core of semantic web-based application systems. The Semantic Web realizes semantic interoperability by utilizing ontology and agent technology, and suggests a new paradigm that can develop existing information expression-oriented web into knowledge-based semantic web.

In addition, there is an attempt to build a knowledge base similar to the semantic web concept by creating a document that is semantically annotated with respect to general documents in addition to web documents. In particular, in order to enable Adobe to annotate PDF documents, XMP, a metadata representation method, has been developed and supported.

In summary, while large web documents or general documents are conventionally written with a human-centered closed and visual arrangement, recently, a large amount of documents can be searched more effectively and efficiently using a machine agent, and furthermore, the knowledge base. In order to construct, by attaching semantic annotations to documents based on ontology with semantic interoperability, not only human but also machine agents can identify and search semantic relations of documents and classify and manage documents. Increasingly, research is underway.

Standardization techniques for semantically weaving web documents or general documents scattered as described above are developed in each field, and RDF, OWL, TopicMaps, KIF, OWL-S, WSDL-S, AJAX, XMP, Dublin Core, Various types of SIOC, FOAF, SKOS, and SCOT are suggested. But what's most important is an integrated tool and system for creating and managing documents, such as semantically structured Webs, or for applying these forms to documents that have already been created.

SUMMARY OF THE INVENTION An object of the present invention is to solve the problems described above, and to provide a semantic annotation system and method for annotating a document using a semantic knowledge base structured by ontology of semantic information of the document.

Another object of the present invention is to subdivide the information of the document by the component of the document rather than by the document, and to annotate using a semantic information knowledge base composed of an ontology that provides semantic interoperability to the document. It is to provide a semantic annotation system and method that enables not only people but also machines to grasp meaning and find suitable results to enable semantic based document search.

Another object of the present invention consists of an ontology capable of securing semantic interoperability with respect to metadata for annotation, and automatically finds an annotation of a document that can be properly annotated through this semantic knowledge base. It is to provide a semantic annotation system and method.

The present invention relates to a semantic annotation system having an ontology repository, the semantic information management unit for storing the semantic information in the ontology repository to build a semantic information knowledge base; A document management unit for importing a document and attaching semantic information to the document; An annotation automatic processing unit for annotating the components of the document with the semantic information when the semantic information related to the elements of the document exists in the semantic information knowledge base; And an annotation preparation unit for annotating the semantic information to the elements of the document with respect to the semantic information designated by the user and the elements of the document.

In addition, the semantic annotation system according to the present invention is characterized in that the user can be annotated before annotating the components of the document automatically with the semantic information.

Further, in the semantic annotation system according to the present invention, the component of the document includes a vocabulary, and the annotation automatic processing unit is matched when the vocabulary of the document and the name attribute (value) of the semantic information instance are the same. Or a morpheme of the vocabulary of the document and a morpheme attribute (value) of the semantic information class.

In addition, in the semantic annotation system according to the present invention, the annotation automatic processing unit generates an instance of the class if the semantic information class having the same morpheme attribute (value) as the morpheme of the vocabulary of the document exists. The attribute value of the name attribute of the instance is set as the vocabulary of the document, and the vocabulary of the document is annotated with the instance.

In addition, in the semantic annotation system according to the present invention, the document management unit attaches only connection information for the instance if the semantic information annotated to the document exists as an instance in the semantic information knowledge base.

In addition, in the semantic annotation system according to the present invention, the document management unit is characterized in that the attached semantic information to the document in any one of the format of XML, XMP, RDF format attached.

In the semantic annotation system according to the present invention, the semantic information management unit loads an ontology designated by a user from the semantic knowledge base stored in the ontology repository, and the annotation automatic processing unit loads the loaded semantic knowledge base. It is characterized by checking whether there is semantic information related to the target.

The semantic annotation system according to the present invention may further include an interface unit configured to transmit a screen including a ontology view showing a semantic information knowledge base, a document view showing a document, and an annotation view showing the annotation information to a user computer. The information may be information on activated semantic information of the ontology view that annotates components of the document activated in the document view.

In the semantic annotation system according to the present invention, the interface unit highlights a component of the document on the screen that the annotation automatic processing unit notifies the component of the document that can be annotated. .

The semantic annotation system according to the present invention may further include a faceted browser that receives a search condition for an attribute, searches for a semantic information instance that meets the search condition, and displays a result. .

In addition, the semantic annotation system according to the present invention, characterized in that it further comprises a semantic information publishing unit for converting and storing the annotations created in the document in the SCOT (Social Semantic Cloud of Tags) format.

In the semantic annotation system according to the present invention, the document is any one of a web document, a PDF document, an image file, a video file, and an audio file, and the components of the document are the entire document, text or vocabulary, images, and voice. , At least one of a video and a link.

In addition, in the semantic annotation system according to the present invention, the document is a document embedded in a user computer or a document that is transferred and connected to the Internet is characterized in that the imported.

The present invention also relates to a semantic annotation system having an ontology repository, comprising: a semantic information management unit for storing semantic information (metadata) in the ontology repository to construct a semantic information knowledge base; A document management unit for importing a document and attaching semantic information to the document; In order to determine the relationship between the component and the semantic information, an attribute is defined in an element of the document, an attribute of semantic information to be compared with an attribute of the element is determined, and a comparison method of the attributes for determining the relationship An automatic processing rule manager to store and manage the automatic processing rule for determining a value; For each component of the document, if the attribute of the semantic information to be compared with the attribute of the component is in the automatic processing rule, the semantic information having the attribute of the semantic information is searched in the semantic information knowledge base, and the retrieved semantic information is retrieved. An annotation automatic processing unit for annotating the components of the document with the retrieved semantic information when it is determined that the attributes (values) of the components and the attributes (values) of the components are related; And an annotation creation unit for annotating the semantic information to the elements of the document with respect to the semantic information designated by the user and the elements of the document.

In the semantic annotation system according to the present invention, a component of the document includes a vocabulary, and the automatic processing rule manager defines the vocabulary itself as a name attribute in the vocabulary, and compares the vocabulary with the name attribute of the vocabulary. A semantic information that defines an attribute of information as a name attribute, defines a rule that the attributes are related, or defines a morpheme included in a vocabulary in the vocabulary as a morpheme attribute and compares the morpheme attribute of the vocabulary. Set the attribute of as a morpheme attribute, and if the attributes are the same, it is characterized in that to store the rules that are determined to be relevant as an automatic processing rule.

The present invention also relates to a semantic annotation method using an ontology repository, comprising: (a) storing semantic information (metadata) in the ontology repository to build a semantic information knowledge base; (b) loading some of the stored semantic knowledge bases by user designation; (c) retrieving a document by user specification; (d) annotating the components of the document with the semantic information when the semantic information related to the components exists in the semantic information knowledge base for all the components of the document; (e) annotating the semantic information to the elements of the document with respect to the semantic information designated by the user and the elements of the document; (f) attaching the semantic information to the document.

Further, in the semantic annotation method according to the present invention, step (d) is characterized by notifying the components of the document that the annotation can be performed before annotating the semantic information.

In the semantic annotation method according to the present invention, the step (d) may include the following steps (d2) to (d5) if the component is a vocabulary for each component of the document (d1). Performing; (d2) If there is an instance of semantic information having the same name attribute (value) as the vocabulary of the document, annotate the vocabulary of the document with the instance, perform step (d5), and if not exist, step (d3) Performing; (d3) if a semantic information class having the same stemming attribute (value) as the morpheme of the document's vocabulary exists, create an instance for the class, and set the attribute value of the name attribute of the created instance as the vocabulary of the document; step; (d4) annotating the vocabulary of the document to the instance; (d5) performing the steps (d2) to (d4) for all components of the document.

In the semantic annotation method according to the present invention, the step (f) is characterized in that only the connection information for the instance is attached if the semantic information annotated to the document exists as an instance in the semantic information knowledge base.

In the semantic annotation method according to the present invention, step (f) is characterized in that the semantic information is attached to the document in one of XML, XMP, and RDF formats.

Further, in the semantic annotation method according to the present invention, the step (d) is characterized by highlighting the components of the document that are notified of the components of the document that can be annotated on the screen.

In addition, the semantic annotation method according to the present invention, characterized in that it further comprises the step of converting and storing the annotations written in the document before the step (f0) (f) in the SCOT (Social Semantic Cloud of Tags) format .

In the semantic annotation method according to the present invention, the document is any one of a web document, a PDF document, an image file, a video file, and an audio file, and the components of the document are the entire document, text or vocabulary, images, and voice. , At least one of a video and a link.

In the semantic annotation method according to the present invention, the document may be a document embedded in a user computer or a document imported by being connected to the Internet.

The present invention also relates to a semantic annotation method using an ontology repository, comprising: (g) constructing a semantic information knowledge base by storing semantic information (metadata) in the ontology repository; (h) In order to determine the relationship between the components of the document and the semantic information, the attributes are defined in the components of the document, the attributes of the semantic information to be compared with the attributes of the components, and the relationship to determine the relationship Creating and storing an automatic processing rule for determining a method of comparing the attributes; (i) loading some of the stored semantic knowledge bases by user designation; (j) retrieving a document by user specification; (k) for each component of the document, if the attribute of the semantic information to be compared with the attribute of the component is in the automatic processing rule, the semantic information having the attribute of the semantic information is searched in the semantic information knowledge base, Annotating the elements of the document with the retrieved semantic information when it is determined that the attributes (values) of the retrieved meaning information are related by comparing the attributes (values) of the elements; (l) annotating the semantic information to components of the document with respect to the semantic information designated by the user and the components of the document; (m) attaching the semantic information to the document.

In the semantic annotation method according to the present invention, the components of the document include a vocabulary, and in step (h), the vocabulary itself is defined as a name attribute in the vocabulary, and the name attribute of the vocabulary is compared with the name attribute of the vocabulary. A semantic information that defines an attribute of semantic information as a name attribute, defines a rule that is related when the attributes are identical, or defines a morpheme included in a vocabulary in the vocabulary as a morpheme attribute and compares the morpheme attribute of the vocabulary. Set the attribute of as a morpheme attribute, and if the attributes are the same, it is characterized in that to store the rules that are determined to be relevant as an automatic processing rule.

The invention also relates to a computer readable recording medium having recorded the semantic annotation method described above.

As described above, according to the semantic annotation system and method according to the present invention, by using the semantic information knowledge base structured the document information is divided into the component units of the document rather than the document unit, and the semantic information of the document is structured by the ontology By annotating, the semantic-based document retrieval can be obtained, in which not only people but also machines can grasp the meaning and find suitable results.

In addition, according to the semantic annotation system and method according to the present invention, by automatically or semi-automatically annotated with semantic information constructed ontologies to a large number of conventional web documents or general documents, it is possible to quickly The effect of annotating correctly is obtained.

In addition, according to the semantic annotation system and method according to the present invention, by supporting a PDF document commonly used as a standard document in addition to the web document, or by supporting a SCOT format commonly used as a standard for meta tags, compatibility with other standard technologies It is possible to obtain an effect of increasing the ease of use.

Prior to describing the details for carrying out the present invention, the concept of a document, semantic information, semantic information knowledge base, and annotations to be described below will be described.

The document according to the present invention refers to an electronic file made to transmit information through a sense of sight or hearing of a person. Therefore, general electronic documents such as PDF files or Korean files, or image files or video files correspond to this. On the other hand, while documents easily convey information to people, machines such as search agents have difficulty in interpreting such document information semantically. The solution to this is metadata.

The semantic information according to the present invention refers to metadata that describes information of a document so that a machine such as a search agent can also interpret semantic information of the document. The semantic information does not mean a metadata for structuring information of a document or a standard standard for indicating metadata. Therefore, standard specifications for indicating metadata such as HTML, XML, and RDF do not correspond to semantic information. However, in the present invention, the standard specification of a known metadata notation is appropriately selected and used to indicate semantic information in a document.

The semantic information knowledge base according to the present invention means the whole semantic information stored in the ontology repository by constructing the semantic information into an ontology. Since the semantic information must be metadata that is created and structured by a certain rule in order for the machine to interpret it like general metadata, the semantic information is composed of ontology, created and structured by a certain rule, and constructed into a semantic information knowledge base. will be. This knowledge base can be used to annotate semantic information in other documents, rather than creating and appending semantic information to one document. If the semantic information is written in a document and is not already in the semantic knowledge base, the semantic information is created and added to the semantic knowledge base. Therefore, the semantic information knowledge base is an ontology repository that continues to evolve by adding, changing, and updating.

 In addition, the annotation according to the present invention means creating and attaching the semantic information to the document.

DETAILED DESCRIPTION Hereinafter, specific contents for carrying out the present invention will be described with reference to the drawings.

In addition, in describing this invention, the same code | symbol is attached | subjected and the repeated description is abbreviate | omitted.

As shown in FIG. 1A, the semantic annotation system 30 according to the present invention is embedded in a computer and used directly by the user 15, or as shown in FIG. 1B, connected to the Internet 5 to the user computer 15. It is also used by connecting.

In the former case, the semantic annotation system 30 may be installed in a computer and used directly by the user 15. The computer may be a general-purpose general purpose computer or a dedicated computer for professionally processing semantic annotations. Since the computer is a known technology commonly used in the art, detailed description thereof will be omitted.

The document 25 may be a document read from the outside or a document inside the computer. The document 25 may be any document that can be processed by a general computer such as a web document or a PDF. However, in the present invention, a web document or a PDF document will be mainly described as an embodiment. Since the above documents are well-known technologies commonly used in the art, detailed description thereof will be omitted. The semantic annotation system 30 reads the document 25 and automatically processes it internally, or attaches an annotation to the document 25 at the command of the user 15. The document 25 may be a document that is not annotated at all, or may be a document that has already been annotated. That is, the semantic annotation system 30 may perform a function such as adding a new annotation to the document 25, modifying or deleting the document.

The ontology repository 40 is a general ontology repository for organizing and storing information as an ontology. The ontology repository 40 may define a class, define a class, define attributes and input attribute values, create and delete instances, and input and change attribute values. It has a normal ontology DB engine that can perform functions. The ontology reservoir 40 is also a well-known technology used in this field, so a detailed description thereof will be omitted.

As shown in FIG. 1B, the semantic annotation system 30 according to the present invention may be connected to the Internet 5 to annotate document information scattered on the Internet. On the Internet 5, it is divided into a service provider server 20 which provides documents to be annotated, and a service user computer 10 using this information. The semantic annotation system 30 functions as a device of a service broker that supports annotation so that the service user 10 can more easily search and use information provided by the service provider 20. In other words, the service broker supports the basic ontology repository or annotation system necessary for annotating documents. Using this, the service provider 20 directly creates an annotation and provides an attached document or a document that can be annotated. The service user 10 often uses the document to obtain information, but can annotate the document directly. That is, the semantic annotation system 30 can use both a service user and a service provider, but can limit the function according to each subject. For example, creation of a new class may be prohibited to a service user or a service provider. This policy is a well-known technique commonly used in the art, so detailed description thereof will be omitted.

Next, a semantic annotation system according to an embodiment of the present invention will be described with reference to FIG. 2.

2 is a block diagram of a configuration of a semantic annotation system according to an embodiment of the present invention.

As shown in FIG. 2, the semantic annotation system 30 according to an embodiment of the present invention includes a semantic information managing unit 31, a document managing unit 32, an annotation automatic processing unit 33, and an annotation preparing unit 34. It is composed. In addition, the semantic annotation system 30 is interlocked with the ontology reservoir 40. In addition, the interface unit 36 may include a semantic information publishing unit 37 and a facet browser 38.

Meanwhile, another embodiment may be configured by adding an automatic processing rule manager 35 to the semantic annotation system 30. However, it will be described below without dividing the description into separate embodiments.

The semantic information manager 31 configures semantic information into an ontology, stores the semantic information in the ontology repository, and constructs the semantic information knowledge base. Semantic information is metadata for semantically describing a document. In other words, a general document is composed of sentences written in natural language, but its meaning is not understood unless the natural language is mechanically interpreted. For example, if a general paper is written as a PDF document, the title, author, summary, text, and comments can be seen by humans, but the machine cannot interpret the meaning. Therefore, the metadata that records the author, the title, and the like for these documents is semantic information. The semantic information may record semantic information about a phrase, a vocabulary, an image, a video, a voice, etc. in the document, in addition to the semantic information of the entire document as described above.

In addition, the semantic information is composed of an ontology for semantic interoperability. As in the above example, the semantic relationship can be established with each other by grouping the title, the author, the summary, and the like into properties of one class. In addition, since semantic information may be configured differently according to types of documents, that is, general books, manuals, reports, dictionaries, etc., the ontology is classified into upper classes, lower classes, and the like to have structural meaning. Configure. Therefore, the semantic knowledge base of the class is formed in a tree form, each class can create an instance, the instances created in the class is managed as a list of the class. It also manages the properties belonging to each class or instance as a list.

In addition, the semantic information may be constructed as one or more semantic information knowledge bases composed of ontology. Ontologies may be classified by type, such as domain ontology, ontology for conceptual expression, and everyday ontology, and various ontologies may be configured according to a field (particularly, domain ontology). In addition, there may be a case where the ontology already configured is used. In this case, ontologies may be constructed with different concepts for the same semantic information. Similarly, the semantic knowledge base may be implemented with various knowledge bases, and knowledge bases of different concepts may be constructed for the same semantic information. Therefore, the ontology store stores a knowledge base of all semantic information, but only a part thereof can be loaded and used according to a user's designation. That is, the semantic information management unit 31 may individually select and load the semantic information knowledge base associated with the document to be annotated. If only a part is designated and loaded, the semantic information to be processed is loaded. Limited to semantic knowledge base.

The document management unit 32 retrieves a document and attaches semantic information to the document. Documents can include all types of documents, just as there are different documents, depending on the software that creates them. The documents preferably have the ability to attach annotations. For example, web documents support ontologies such as RDF, XML, SCOT, meta tags, etc., and thus have a platform for attaching annotation information to web documents. PDF also has an XMP platform, which allows users to edit and search semantic information. The document management unit 32 attaches semantic information in XMP format when the document is a PDF document. In addition, it can be attached in the form of XML, XMP, or RDF. However, Korean documents, word documents, image files, video files, etc. do not have such a function, so to attach an annotation, a means for processing them separately is required. For example, annotation information performed on the document can be separately managed and only the document and connection information can be provided. Nevertheless, it is not easy to deal with annotations on vocabulary and phrases or descriptions (of image files or video files) at specific locations in the document.

Meanwhile, the document manager 32 attaches semantic information that is annotated to the document. However, if the semantic information annotated to the document exists as an instance in the semantic information knowledge base, it is preferable to attach only the connection information for the instance. That is, the document management unit 32 may attach only the connection information for the instance if the semantic information annotated to the document exists as an instance in the semantic information knowledge base. This is because attaching the entire instance including ontology information increases the size of the document. However, if you attach a rule by attaching only the required attribute values among instances, the size may be larger than attaching only the connection information, but it may be faster when searching documents later. One example is

Meanwhile, as described above, the document loaded from the document manager 32 may be a document embedded in a user computer or a document imported by being connected to the Internet. In other words, it does not matter where the document is stored.

The annotation automatic processing unit 33 annotates the semantic information to the component of the document when the semantic information related to the component of the document exists in the semantic information knowledge base.

Generally, a document is composed of an entire document, pages, text or vocabulary, images, voice, video, links, and the like. Such elements will be referred to as document components. A typical document may include most of the above components, but a document file such as an image or a video may include only an image or a video. The whole document is about the document itself. Text refers to what is described as text in a document, which is a common natural language description. The text may be further subdivided into paragraphs, phrases, vocabulary, etc., and viewed as elements of the document. Paragraphs are divided into headings, summaries, texts, and comments, depending on the type. In other words, a document is composed of paragraphs, and its components may be regarded as types of summary, text, comments, etc., and documents may be regarded as summary paragraphs, text paragraphs, and comment paragraphs. have. Also, images, voices, videos, tables, links, etc. are specially shaped elements other than text. Special form elements other than text may be designated as elements of one document, and each voice, video, etc. may be classified into types.

That is, the present invention can be carried out by dividing the various components as described above in each embodiment. The reason for classifying the components of the document as described above is that annotations can be added by adding semantic information according to the respective components, and such annotations can enrich the search.

On the other hand, the components of a document have their own characteristics. For example, the characteristics of the entire document may include a title, author, creation date, document type, size, and the like. The characteristics of a paragraph may include a type, size, order, and the like. , Prefix, suffix, morpheme, and so on. In the present invention, the characteristics of the components of the document as described above will be referred to as attributes of the components of the document. In particular, the vocabulary itself among the characteristics of the vocabulary will be referred to as the "name attribute" of the vocabulary. For example, the attribute value of the name attribute of the "sea" vocabulary is "sea". That is, the word "sea" has features such as the word "sea", two letters, nouns, etc., which means that the property of the word "sea" is called "sea" in the present invention.

On the other hand, the determination of whether the components of the document and the semantic information is related is determined by comparing the attributes of the components and the attributes of the semantic information. For example, the attributes may be the same, which attributes belong to different attributes, or the like. In addition, if there is similarity according to the similarity judgment used in natural language search or the like, it may be determined that both attributes are related. Semantic information consists of ontology, so it is a class or an instance. Since the class or instance has an attribute, the attribute of the class or instance is called an attribute of semantic information. More precisely, when semantic information is an instance, an attribute of semantic information means an attribute of semantic information instance.

The annotation automatic processor 33 checks the attribute of the component to automatically annotate the components of the document. For example, when the term "banana" appears, when annotating automatically with the "banana" instance of semantic information, the name attribute of the "banana" vocabulary is "banana", and the name attribute of the "banana" instance is "banana". Since the name attribute of the vocabulary and the name attribute of the semantic information are the same, the semantic information may be annotated.

If the semantic information class having the same morpheme attribute (value) as the morpheme of the vocabulary of the document exists, the annotation automatic processing unit 33 creates an instance for the class, and the attribute value of the name attribute of the generated instance. Is set to the vocabulary of the document, and the instance is annotated to the vocabulary of the document. For example, "Seoul National University" is a vocabulary for schools, so if you have a stemming of "school", you can create an instance of "Seoul National University" in the class "School" and annotate this instance automatically. have. In this case, if a morpheme attribute is defined in the "school" semantic information class and "school" is defined in the morpheme attribute value, "school" appears in the morpheme attribute of the vocabulary, and a class having the morpheme attribute as "school" exists. Just annotate. At this time, the name attribute of the instance is automatically entered as "Seoul National University". "Seoul National University" is also a name attribute of "Seoul National University" vocabulary.

Representative methods for automatically annotating the vocabulary of the document by the annotation automatic processing unit 33 include the above two methods of matching by name and morpheme. However, when the two methods are used together, a search is first performed on whether a name matches, and when there is no semantic information corresponding to the name, a morpheme matching method should be applied. This is because if you have a name, you just annotate it, and you don't have to do stemming.

Another embodiment of the annotation automatic processing unit 33 according to the present invention will be described. Although the annotation automatic processing unit 33 has previously described an embodiment of automatically annotating semantic information, there may be another embodiment in which an annotation is semi-automatically performed by a user's command without automatically annotating.

That is, the annotation automatic processing unit 33 notifies that the semantic information can be annotated to the component of the document when the semantic information related to the component of the document exists in the semantic information knowledge base. The annotation automatic processing unit 33 does not automatically, but notifies the user of only the vocabulary that can be annotated and allows the user to check whether the annotation is present. The method of finding the semantic information related to the elements of the document by the annotation automatic processing unit 33 is the same as the above-described embodiment. However, it does not annotate, and in particular, does not create an instance of that class when comparing morphemes to determine if they are related.

Meanwhile, both embodiments of the annotation automatic processing unit 33 according to the present invention may be implemented in the semantic annotation system 30 in a form in which a user performs one of two options by selection. .

The annotation preparing unit 34 annotates the semantic information to the elements of the document with respect to the semantic information designated by the user and the elements of the document. The semantic annotation system 30 according to an embodiment of the present invention supports a user's direct annotation as a tool for conveniently supporting the annotation. That is, the user may select an element of the document and find the semantic information to be annotated in the semantic information knowledge base and annotate the element. At this time, if necessary, an instance may be created, attribute values required for the instance may be input, and the instance may be annotated. As described above, the components of the document can be selected from the entire document, text, vocabulary, voice, video, and image.

The following is related to a semantic annotation system according to another embodiment of the present invention, and is an embodiment in which an automatic processing function is extended by separately managing or adding or modifying a rule capable of automatic processing.

The automatic processing rule manager 35 defines an attribute in an element of the document, determines an attribute of semantic information to compare with an attribute of the element, in order to determine whether the element of the document is related to semantic information, It stores and manages automatic processing rules that determine how to compare the attributes to determine the relationship. And the annotation automatic processing unit 33, for each component of the document, if the attribute of the semantic information compared with the attribute of the component is in the automatic processing rule, the semantic information having the attribute of the semantic information is the semantic information. If it is determined that the attribute (value) of the semantic information retrieved by searching in the knowledge base and the attribute (value) of the component is related and annotated, the component of the document is annotated with the retrieved semantic information.

That is, in the expanded embodiment, the other elements are the same, the automatic processing rule manager 35 is added, and the function of the automatic annotation processor 33 is changed.

The automatic processing rule manager 35 allows to define the attributes of the components of the document described above. For example, the name attribute of a vocabulary may define the vocabulary itself as a name attribute in the vocabulary. In addition, the morpheme attribute of a vocabulary can be defined as the morpheme which a vocabulary has. This uses a conventional lexical analyzer. Since it is known in the art, specific description thereof will be omitted. An attribute of semantic information to be compared with an attribute of the vocabulary is defined, and a comparison operator is defined. For example, one of the comparison operators is selected to be the same, belong, include, large, small, or equal. Therefore, the automatic processing rule is a rule consisting of a definition of an attribute of a document element, an attribute of semantic information, and a comparison operator.

For example, the automatic processing rule manager 35 sets the vocabulary itself as a name attribute in a vocabulary that is a component of a document, and sets an automatic processing rule to set an attribute of semantic information to be compared with the name attribute as a name attribute, or In the vocabulary, a morpheme of the vocabulary is defined as a morpheme attribute, and an automatic processing rule for setting an attribute of semantic information compared with the morpheme attribute as a morpheme attribute is created and stored.

As described above, the use of automatic processing rules can continue to improve the automatic processing capacity by discovering additional processing rules. For example, if there is an explanatory attribute of an image that does not appear on the document, the vocabulary included in the description may be automatically processed and annotated. In addition, in the case of a paragraph, if the type of document is a paper or the like, a rule that can automatically process whether the paragraph, order, length, or position of the title, author, summary, etc. may be generated.

The interface unit 36 transmits a screen composed of an ontology view showing a semantic information knowledge base, a document view showing a document, and an annotation view showing the annotation information to the user computer. In addition, the interface unit 36 highlights a component of the document on the screen that the annotation automatic processing unit notifies the component of the document that can be annotated.

3 is a diagram illustrating a screen configuration of a semantic annotation system according to an embodiment of the present invention. As shown in FIG. 3, the screen includes an ontology view showing a semantic information knowledge base on the left side, a document view showing a document on the upper right side, and an annotation view showing the annotation information on the lower right side. The ontology view is divided into the bottom and the top, and the above shows the class in tree form, and the bottom shows the instance or property in list form. The document view has a wider view because it shows the document. The annotation view shows the attributes of the semantic information. In particular, the annotation view shows information related to the document activated in the document view and semantic information activated in the ontology view.

The semantic information publishing unit 37 converts and stores the annotations created in the document in SCOT (Social Semantic Cloud of Tags) format.

SCOT (Social Semantic Cloud of Tags) is an ontology for expressing tags used in online community. It structurally expresses tags in RDF and defines tags semantically. This allows users to reuse tags or share them with other users, creating an environment where people, content, and tags are semantically connected to social networks. SCOT Ontology is connected to FOAF, SIOC, and SKOS, and all activities to create and share SCOT Ontology are done in the Semantic Web environment.

The semantic annotation system according to an embodiment of the present invention converts and stores annotations in the SCOT format, which is the most representative tag ontology, for compatibility with tag ontology.

The facet browser 38 searches for semantic information and annotations by determining a matching value for at least one of a class, an instance, and an attribute. Faceted browsing is a method of searching for each component of ontology using the characteristics of ontology. For example, in order to perform a "movie search starring Gam Woo Sung", a class of a person is specified and the condition that the name attribute is "Gam Woo Sung" is found to find the "Gam Woo Sung" instance, and the actor attribute is "Gam Woo Sung" instance in the movie class. Perform a search by finding a movie instance.

Accordingly, the facet browser 38 searches based on a condition for determining a class target to search for, a facet for determining whether to match, an instance that is an actual value of the attribute, or a comparison value that is an attribute value. . The search results are instances or documents of semantic information. In the case of a complex search, the sequential search process is recorded.

In summary, the faceted browser receives a search condition for an attribute, searches for a semantic information instance that meets the search condition, and displays a result. That is, the main search is a search by attributes, and classes and the like can be defined to further limit the search.

Next, a screen shown in the semantic annotation system according to an embodiment of the present invention will be described with reference to FIGS. 4 to 7.

4 is a diagram illustrating a screen for loading a semantic information knowledge base according to an embodiment of the present invention. As shown in Fig. 4, when the knowledge base is loaded, the semantic information class tree in the form of a tree is shown.

5 is a diagram illustrating a screen for loading a PDF document according to an embodiment of the present invention. As shown in FIG. 5, when a PDF document is loaded, the document is displayed on the upper right side of the screen.

6 is a diagram illustrating a screen for automatic annotation processing according to an embodiment of the present invention. As shown in FIG. 6A, when annotated automatically, components of the document annotated in the document view are activated, and attributes of semantic information are shown in the annotation view at the bottom. In addition, as shown in Figure 6b, the components of the semi-automatically annotable document are highlighted in yellow.

7 is a diagram illustrating a screen for facet browsing according to an embodiment of the present invention. As shown in FIG. 7A, a Person class is selected in a type, a name is selected in a facet, and a "feeling" instance is displayed by selecting a "feeling" name appearing in the list. The result is shown on the lower information screen. As shown in FIG. 7B, the movie class is selected, and the name attribute is selected for the facet. At this time, the "Gam Woo Sung" instance searched just before the search process (history) finds the results of the movie instances starring "Gam Woo Sung" as a condition. The result is shown in the information below.

Next, a semantic annotation method according to an embodiment of the present invention will be described with reference to FIGS. 8 and 9. FIG. 8 is a flowchart illustrating a semantic annotation method according to an embodiment of the present invention, and FIG. 9 is a flowchart illustrating a method of annotating semantic information to a vocabulary of a document according to an embodiment of the present invention.

As shown in Figure 8, the semantic annotation method according to an embodiment of the present invention comprises the steps of (a) storing the semantic information (metadata) in the ontology repository to build a semantic information knowledge base (S10); (b) loading some of the stored semantic knowledge bases by user designation (S20); (c) step S30 of retrieving a document by user designation; (d) annotating the components of the document with the semantic information when the semantic information related to the components exists in the semantic information knowledge base for all the components of the document (S40); (e) annotating the semantic information with respect to the semantic information designated by the user and the elements of the document (S50); (f) the semantic information is attached to the document (S60).

As shown in FIG. 9, in the step (d), for each component of the document, if the component is a vocabulary, the steps (d2) to (d5) are performed (S41). ); (d2) If there is a semantic information instance having the same name attribute (value) as the vocabulary of the document, annotate the vocabulary of the document with the instance, perform step (d5), and if it does not exist, step (d3) Performing step (S42); (d3) if a semantic information class having the same stemming attribute (value) as the morpheme of the document's vocabulary exists, create an instance for the class, and set the attribute value of the name attribute of the created instance as the vocabulary of the document; Step S43; (d4) annotating the vocabulary of the document with the instance (S44); (d5) For all the components of the document, it is further subdivided into the step (S45) of performing the steps (d2) to (d4).

In step (d), the component of the document is notified before the semantic information is annotated (S41).

In step (f), if the semantic information annotated to the document exists as an instance in the semantic information knowledge base, only the connection information for the instance is attached.

In step (f), the semantic information is attached to the document in one of XML, XMP, and RDF formats.

In step (d), the components of the document notified to the components of the document that can be annotated are highlighted on the screen.

The semantic annotation method further includes converting an annotation written in the document before step (f0) and (f) into a SCOT (Social Semantic Cloud of Tags) format and storing the converted annotation.

The document is any one of a web document, a PDF document, an image file, a video file, and an audio file, and a component of the document includes at least one of a whole document, text or vocabulary, an image, an audio, a video, and a link. .

The document is a document embedded in a user's computer or a document imported and connected to the Internet.

The step (b) and the step (c) may be performed simultaneously or the progress of the posterity relationship may be changed.

Next, a semantic annotation method according to other embodiments of the present invention will be described with reference to FIG. 10. 10 is a flowchart illustrating a semantic annotation method according to another embodiment of the present invention.

As shown in FIG. 10, the method relates to a semantic annotation method using an ontology repository, comprising: (g) constructing a semantic information knowledge base by storing semantic information (metadata) in the ontology repository (S210); (h) In order to determine the relationship between the components of the document and the semantic information, the attributes are defined in the components of the document, the attributes of the semantic information to be compared with the attributes of the components, and the relationship to determine the relationship Creating and storing an automatic processing rule for determining a method of comparing the attributes (S215); (i) loading some of the stored semantic knowledge bases by user designation (S22); (j) retrieving a document by user specification (S230); (k) For each component of the document, if the attribute of the semantic information comparing with the attribute of the component is in the automatic processing rule, the semantic information having the attribute of the semantic information is searched and searched in the semantic information knowledge base. Comparing the attribute (value) of the semantic information with the attribute (value) of the component and annotating the element of the document with the retrieved semantic information (S240); (l) annotating the semantic information with the semantic information designated by the user and the elements of the document (S250); (m) the method is attached to the document (S260).

The component of the document includes a vocabulary, and in step (h), the vocabulary itself is defined as a name attribute in the vocabulary, the attribute of semantic information comparing with the name attribute of the vocabulary is defined as a name attribute, and the attribute If they are the same, the rule judged to be related or the morpheme included in the vocabulary in the vocabulary are defined as the morpheme attribute, the attribute of the semantic information comparing with the morpheme attribute of the vocabulary is determined as the morpheme attribute, and the attributes are the same. If you do so, create a rule that you consider to be related as an automatic processing rule and save it.

For the insufficient description of the semantic annotation method, refer to the description of the semantic annotation system described above.

As mentioned above, although the invention made by this inventor was demonstrated concretely according to the said Example, this invention is not limited to the said Example and can be variously changed in the range which does not deviate from the summary.

The present invention is applicable to the field of searching a large number of documents, and in particular, for documents commonly used on the Internet such as web documents or PDFs, a search based on semantic interoperability besides a simple keyword search is performed to find a desired document effectively. Applicable to the field.

In addition, the present invention can be applied to the field of building a whole knowledge base by classifying and annotating documents based on semantic interoperability.

FIG. 1A is a diagram illustrating a configuration of an entire system for implementing a semantic annotation system and method according to the present invention with a built-in PC.

1B is a diagram illustrating the configuration of an entire system for implementing a semantic annotation system and method in a network according to the present invention.

2 is a block diagram of a configuration of a semantic annotation system according to an embodiment of the present invention.

3 is a diagram illustrating a screen configuration of a semantic annotation system according to an embodiment of the present invention.

4 is a diagram illustrating a screen for loading a semantic information knowledge base according to an embodiment of the present invention.

5 is a diagram illustrating a screen for loading a PDF document according to an embodiment of the present invention.

6 is a diagram illustrating a screen for automatic annotation processing according to an embodiment of the present invention.

7 is a diagram illustrating a screen for performing a multi-faceted search according to an embodiment of the present invention.

8 is a flowchart illustrating a semantic annotation method according to an embodiment of the present invention.

9 is a flowchart illustrating a method of annotating metadata of a vocabulary of a document according to an embodiment of the present invention.

10 is a flowchart illustrating a semantic annotation method according to a second embodiment of the present invention.

Explanation of symbols on the main parts of the drawings

5: internet 10: user terminal

20: Document Server 25: Document

30: semantic annotation system 40: ontology DB

31: semantic information management unit 32: document management unit

33: Annotation automatic processing unit 34: Annotation creation unit

35: automatic processing rule management unit 36: interface unit

37: Semantic Information Publishing Department

Claims (27)

A semantic annotation system having an ontology repository, A semantic information manager to store semantic information in the ontology repository to construct a semantic information knowledge base; A document management unit for importing a document and attaching semantic information to the document; An annotation automatic processing unit for annotating the components of the document with the semantic information when the semantic information related to the elements of the document exists in the semantic information knowledge base; An annotation creation unit for annotating the semantic information to the elements of the document with respect to the semantic information designated by the user and the components of the document; A semantic annotation system comprising a. The method of claim 1, wherein the annotation automatic processing unit, And a semantic annotation system, wherein the semantic annotation system is notified to the user of the document before automatically annotating the semantic information. The method of claim 1, The components of the document include a vocabulary, The annotation automatic processing unit, Determine that the vocabulary of the document and the name attribute (value) of the semantic information instance are the same, or And a morpheme of the vocabulary of the document and a morpheme attribute (value) of the semantic information class are determined to be related. The method of claim 3, wherein the annotation automatic processing unit, If there is a semantic information class having the same morpheme attribute (value) as the morpheme of the document's vocabulary, an instance is created for the class, and the attribute value of the name attribute of the generated instance is defined as the vocabulary of the document. A semantic annotation system for annotating a vocabulary of a document with the instance. The method of claim 1, wherein the document management unit, If the semantic information annotated to the document exists as an instance in the semantic information knowledge base, the semantic annotation system, characterized in that attached only the connection information for the instance. The method of claim 1, wherein the document management unit, Semantic annotation system characterized in that the semantic information in the form of XML, XMP, RDF format attached to the document attached. The method of claim 1, The semantic information management unit loads the knowledge base designated by the user from the semantic information knowledge base stored in the ontology repository, And the annotation automatic processing unit checks whether there is semantic information related to only the loaded semantic information knowledge base. The method of claim 1, It further includes an interface for transmitting a screen consisting of an ontology view showing a semantic information knowledge base, a document view showing a document, an annotation view showing an annotation information to the user computer, And the annotation information is information about activated semantic information of the ontology view that annotates a component of a document activated in the document view. The method of claim 8, wherein the interface unit, And the component of the document that the annotation automatic processor notifies the component of the document that can be annotated on the screen. The method of claim 1, The semantic annotation system of claim 1, further comprising: a faceted browser that receives a search condition for an attribute, searches for a semantic information instance that meets the search condition, and displays a result. The method of claim 1, The semantic annotation system further comprises a semantic information publishing unit for converting and storing the semantic information created in the document in the SCOT (Social Semantic Cloud of Tags) format. The method of claim 1, The document is any one of a web document, PDF document, image file, video file, audio file, The semantic annotation system of claim 1, wherein a component of the document includes at least one of a document, a text or a vocabulary, an image, a voice, a video, and a link. The method of claim 1, The document is a semantic annotation system, characterized in that the document embedded in the user's computer or a document that is imported and connected to the Internet. A semantic annotation system having an ontology repository, A semantic information manager to store semantic information (metadata) in the ontology repository to build a semantic information knowledge base; A document management unit for importing a document and attaching semantic information to the document; In order to determine the relationship between the component and the semantic information, an attribute is defined in an element of the document, an attribute of semantic information to be compared with an attribute of the element is determined, and a comparison method of the attributes for determining the relationship An automatic processing rule manager to store and manage the automatic processing rule for determining a value; For each component of the document, if the attribute of the semantic information to be compared with the attribute of the component is in the automatic processing rule, the semantic information having the attribute of the semantic information is searched in the semantic information knowledge base, and the retrieved semantic information is retrieved. An annotation automatic processing unit for annotating the components of the document with the retrieved semantic information when it is determined that the attributes (values) of the components and the attributes (values) of the components are related; An annotation creation unit for annotating the semantic information to the elements of the document with respect to the semantic information designated by the user and the components of the document; A semantic annotation system comprising a. The method of claim 14, The components of the document include a vocabulary, The automatic processing rule management unit defines the vocabulary itself as a name attribute in the vocabulary, and sets the attribute of the semantic information comparing with the name attribute of the vocabulary as the name attribute, and if the attributes are the same, the rule to determine that the relationship; Alternatively, the morpheme included in the vocabulary is defined as a morpheme attribute in the vocabulary, the attribute of semantic information comparing with the morpheme attribute of the vocabulary is defined as a morpheme attribute, and if the attributes are the same, a rule is determined to be related. Semantic annotation system, characterized in that stored as a rule. In semantic annotation method using ontology repository, (a) storing semantic information (metadata) in the ontology repository to build a semantic information knowledge base; (b) loading some of the stored semantic knowledge bases by user designation; (c) retrieving a document by user specification; (d) annotating the components of the document with the semantic information when all the components of the document have semantic information related to the components in the semantic information knowledge base; (e) annotating the semantic information to the elements of the document with respect to the semantic information designated by the user and the elements of the document; (f) attaching the semantic information to a document; Semantic annotation method comprising a. The method of claim 16, wherein step (d) And informing the components of the document that the semantic information can be annotated before annotating the semantic information. The method of claim 16, wherein step (d) (d1) for each component of the document, if the component is a vocabulary, performing steps (d2) to (d5); (d2) If there is an instance of semantic information having the same name attribute (value) as the vocabulary of the document, annotate the vocabulary of the document with the instance, perform step (d5), and if not exist, step (d3) Performing; (d3) if a semantic information class having the same stemming attribute (value) as the morpheme of the document's vocabulary exists, create an instance for the class, and set the attribute value of the name attribute of the created instance as the vocabulary of the document; step; (d4) annotating the vocabulary of the document to the instance; (d5) performing steps (d2) to (d4) for all components of the document; Semantic annotation method comprising a. The method of claim 16, wherein step (f) comprises: If semantic information annotated to a document exists as an instance in the semantic information knowledge base, the semantic annotation method is characterized by attaching only the connection information for the instance. The method of claim 16, wherein step (f) comprises: And attaching semantic information to the document in one of XML, XMP, and RDF formats. The method of claim 17, wherein step (d), The semantic annotation method of highlighting on the screen the components of the document informing the components of the document that can be annotated. The method of claim 16, (f0) The semantic annotation method further comprises converting and storing semantic information created in the document before the step (f) into a SCOT (Social Semantic Cloud of Tags) format. The method of claim 16, The document is any one of a web document, PDF document, image file, video file, audio file, The component of the document includes at least one or more of the entire document, text or vocabulary, images, voice, video, links. The method of claim 16, The document is a document embedded in a user computer or a semantic annotation method, characterized in that the document is transferred to the Internet imported. In semantic annotation method using ontology repository, (g) storing semantic information (metadata) in the ontology repository to build a semantic information knowledge base; (h) In order to determine the relationship between the components of the document and the semantic information, the attributes are defined in the components of the document, the attributes of the semantic information to be compared with the attributes of the components, and the relationship to determine the relationship Creating and storing an automatic processing rule for determining a method of comparing the attributes; (i) loading some of the stored semantic knowledge bases by user designation; (j) retrieving a document by user specification; (k) for each component of the document, if the attribute of the semantic information to be compared with the attribute of the component is in the automatic processing rule, the semantic information having the attribute of the semantic information is searched in the semantic information knowledge base, If it is determined that the attribute (value) of the retrieved semantic information and the attribute (value) of the component are related, annotating the retrieved semantic information to the component of the document; (l) annotating the semantic information to components of the document, for the semantic information designated by the user and the components of the document; (m) attaching the semantic information to a document; Semantic annotation method comprising a. The method of claim 25, The components of the document include a vocabulary, In step (h), A rule for defining the vocabulary itself as a name attribute in the vocabulary, defining an attribute of semantic information comparing with the name attribute of the vocabulary as a name attribute, and determining that the attributes are related if they are identical; Alternatively, the morpheme included in the vocabulary is defined as a morpheme attribute in the vocabulary, and an attribute of semantic information comparing with the morpheme attribute of the vocabulary is defined as a morpheme attribute, and if the attributes are the same, a rule is determined to be related. Semantic annotation method, characterized in that stored as a rule. A computer-readable recording medium having recorded the semantic annotation method of claim 16.

Images