KR101072147B1 - Method and system for transforming blog posting to Ontology-based information - Google Patents

Abstract

The present invention relates to a method and system for converting a blog post into ontology-based information, and recommends a tag for a post (post) of the blog and annotates based on a tag selected by a user to convert into ontology-based information. Provided are methods and systems for converting blog posts into ontology-based information. The present invention for this purpose comprises the input step of the user terminal inputs the body of the blog and the category for the body via a web browser; A tag recommendation step of searching for an object name of the input text in an ontology of interest corresponding to the category and providing the user name as a plurality of recommendation tags to the user terminal; A selection step of checking, by the user terminal, the recommended plurality of tags and selecting one of the tags; And converting the inputted text into ontology-based information based on the selected recommendation tag and storing the ontology-based information. According to the above configuration, the present invention enables accurate semantic annotation, thereby facilitating next-generation semantic search such as searching, instance analysis, and analysis between instances by instance classification and tags.

Blog, post, ontology, instance, tag

Description

Method and system for transforming blog posting to Ontology-based information}

The present invention relates to a method and system for converting a blog post into ontology-based information, and more particularly, based on information on a topic and an event selected by a user for a tag for a post (post) of the blog. The present invention relates to a method and system for converting a blog post into ontology-based information that can be annotated based on a selected and selected tag and converted into ontology-based information.

The Semantic Web is a next-generation web that became famous for the suggestion of Tim Bernas Lee in 1998. In the distributed environment such as the Internet, it expresses information about resources such as web documents, various files, and services and relationship-meaning information between resources in the form of ontology that can be processed by a machine. Is a framework technology that allows for processing.

The semantic web framework can be divided into ontology and RDF (Resource description) layer that plays an intermediate role for communication between machine and human. The simplest form, RDF, is a triple form of <Subject, Predicate, Object>. Express the concept.

As a simple example, if a person thinks that "banana is yellow" in triple that the machine can understand, it can be expressed as <S: banana, P: color, O: yellow>. By interpreting the triples expressed in this way, the concept of S: banana can interpret and process the concept of having O: color of P: yellow.

Ontology is also knowledge of a particular field, expressed in a form that a machine can understand. Ontology is described in clear logic language, and unlike human language, there is no room for ambiguity or confusion, and it is possible to reason by logic. Due to these features, the semantic web is in the spotlight as the next generation web that can communicate the meaning of machines and people.

As such, the semantic web has chosen a method of expressing human knowledge in a form that can be understood by a machine for meaning communication between people and machines.

However, such a conventional method of implementing the semantic web is not only very difficult to model knowledge of a specific domain, but also annotated by the model still has the possibility of inaccuracy different from the actual meaning. have.

On the other hand, blogs have recently become a form of web publishing paradigm that has become popular with many users. Currently, most blog servers provide an interface for inputting category classification information and tag information so that users can add more detailed semantic information to a written article. In other words, blog users can enter keywords related to the topic of their posts or events they experienced while posting, as tag information, and select the desired category from the category information previously categorized by the system. have.

However, the conventional blog server has no restrictions on tag input, so that in addition to the words that imply a topic, irregular numbers, symbols, and even long sentences can be input. It is mixed with meaning. That is, the current user tag information does not have accurate semantic information attached to it, so it is not easy to identify semantic hierarchical relations or inclusion relations between tags, and as user's unrefined tag information increases, user tag information is searched for articles. And will not be fully utilized in such work as classification. In addition, it is difficult to fully utilize the various resources of the blog system.

The present invention has been proposed to solve the above problems, it is possible for the user to select the recommended tag in relation to the topic or event of the blog post on the ontology-based blog post that can be converted into ontology-based information more accurately It is an object of the present invention to provide a method and system for converting the information into information.

The present invention for achieving the above object is an input step of the user terminal inputs the body of the blog and the category for the body through a web browser; A tag recommendation step of searching for an entity name of the input text in an ontology of interest corresponding to the category and providing the recommendation tag to the user terminal as a plurality of recommendation tags; A selection step of checking, by the user terminal, the recommended plurality of tags and selecting one of the tags; And converting the inputted text into ontology-based information based on the selected recommendation tag and storing the ontology-based information.

Preferably, the tag recommendation step includes: a candidate extraction step of recognizing the entity name by analyzing the input text, and extracting the entity name as a tag candidate; and searching for whether the extracted entity name exists in the ontology of interest; And providing the searched entity name as the recommendation tag along with the searched attribute.

Preferably, in the providing step, the recommendation tag may be compared with the word frequency and the attribute information to rank only the top N items.

Preferably the providing step may provide direct input as an attribute.

Advantageously, the retrieving step may retrieve an instance of the ontology of interest.

Preferably, the converting step may convert the input text into ontology-based information according to the selected tag and its attributes.

Preferably, in the converting step, if there is no selection of the user terminal, the input text may be converted into ontology based information according to a default recommendation tag and its attributes.

Preferably, in the converting step, the ontology-based information may be configured as an OWL file.

According to another aspect of the present invention, a system for converting ontology-based information on a blog includes: a user terminal accessing a blog through a web browser and inputting a text and a category thereof; A resource description framework (RDF) store in which an instance of the ontology of interest corresponding to the category is stored; A blog that searches for the entity name of the input text in the RDF store and provides the recommendation tag to the user terminal as a plurality of recommendation tags, and converts the input text into ontology based information based on the recommendation tag selected in the user terminal. Server; characterized in that it comprises a.

Preferably, the blog server ranks the recommendation tag and provides it to the user terminal along with an attribute retrieved from the RDF store, and provides the recommendation tag selected by the user terminal and its attribute to the conversion unit. A tag attaching unit for providing a default recommendation tag and its attributes to the conversion unit if there is no selection, and converting the input text into the ontology-based information according to any one of the selected recommendation tag and its attributes and the default tag and its attributes. And a storage unit for storing the converted ontology-based information.

Preferably, the tag attaching unit analyzes the input text and extracts an entity name, and ranks the recommendation tag by comparing the word frequency and the attribute information to provide only the top N items as recommendation tags. It may include a tag recommendation unit for providing a recommendation tag selected by the user terminal to the conversion unit, and a search unit for searching whether the extracted entity name is an instance of the ontology of interest of the RDF store.

Preferably, the conversion unit may generate the ontology-based information into an OWL file.

Preferably, the storage unit may include an OWL file storage unit in which the OWL file is stored, and a DBMS (DataBase Management System) in which the body of the blog is stored.

Preferably, the RDF store may have a web communication function.

Preferably, the recommendation tag may include direct input as an attribute.

The method and system for converting a blog post into ontology-based information according to the present invention generate a tag and specific additional information related to the topic or event of a blog post as an ontology instance through feedback with the user, thereby saving time and money as well as the user. By enabling accurate semantic annotations, there is an effect of facilitating next-generation semantic search such as search, instance analysis, and analysis between instances by instance classification and tag.

The present invention creates an instance of a new blog ontology by annotating on the subject of what the contents of the blog post is about and the events that experienced the subject so that the machine can understand and classify the user tag information on the blog post. It is.

In addition, the present invention provides a guide through the feedback of the tag recommendation to the user at the time of input of the tag to perform the semantic annotation linked to the ontology and blog post, allowing the user to directly enter the tag to accurately enter Because semantic annotations are possible, the cost of manual annotations and the inaccuracies of automatic annotations can be overcome simultaneously.

These blog posts can be translated into the OWL language and used as a resource for search, analysis, and reasoning.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a block diagram schematically illustrating a system for converting a blog post into ontology-based information according to an embodiment of the present invention, and FIG. 2 is a block diagram illustrating a detailed configuration of the blog system of FIG. 1.

The blog post to ontology-based information converting system 10 is connected to the blog server 300 to the user terminal 100 to create a blog post, and to recommend the tag for the blog to provide to the user terminal 100 It includes a blog server 300, and the RDF store 400, the ontology is stored.

The user terminal 100 accesses the blog server 300 through the web browser 200, creates a blog body, and inputs a category and a recommended tag for the body provided from the blog server 300 when a tag is set.

The blog server 300 is a tag attachment unit 310 to search the entity name of the input text in the RDF store 400 and provide it to the user terminal 100 as a plurality of recommendation tag, and the recommendation selected in the user terminal 100 A converter 320 converts the text input based on the tag into ontology-based information, and a storage unit 330 in which the converted ontology-based information is stored.

Semantic  Tag Suggestions

The tag attaching unit 310 may include an entity name recognition unit 312 for extracting an entity name by analyzing the input text, a tag recommendation unit 314 for ranking the recommendation tags and providing them to the user terminal 100, and an RDF. The store 400 includes a searcher 316 for searching for a recommendation tag and its attributes.

The entity name recognition unit 312 extracts an entity name for searching for a tag candidate by analyzing the input text. It does not target all nouns contained in the text and extracts the entity names that are expected to exist as ontology instances.

That is, the entity name recognition unit 312 extracts an entity name that may be a candidate of a semantic tag, for example, a book title, a travel destination, a product name of an IT device, and the like from the text.

The tag recommender 314 requests the searcher 316 to search the RDF store 400 for the extracted entity name, ranks the plurality of searched tags with frequency and attribute information, and recommends only the top N recommended tags. As provided to the user terminal 100 as.

Here, the property of the recommendation tag provided to the user terminal 100 preferably includes a direct input.

The tag recommendation unit 314 recommends a tag for an entity name existing as an instance of the ontology of interest so that the entity name to be used as the tag is linked with the ontology.

In addition, the tag recommendation unit 314 provides the recommendation tag selected by the user terminal 100 and its attributes to the conversion unit 320. If there is no selection of the user terminal 100, the tag recommendation unit 314 converts the default recommendation tag and its attributes to the conversion unit ( 320).

The search unit 316 searches whether the entity name extracted by the entity name recognition unit 312 exists in the ontology of interest of the RDF store 400. That is, the search unit 316 searches whether the entity name is an instance of the ontology of interest.

Create a file

The conversion unit 320 converts the input text based on the recommendation tag and its attributes selected by the user terminal 100 into ontology-based information. If there is no selection of the user terminal 100, the default tag and its attributes set to the highest priority are selected. Converts the text entered into ontology-based information.

Here, the ontology-based information is generated in the form of an OWL file. More specifically, the ontology-based information is generated in the form of an OWL file based on a post written by a user and a selected tag. An example of the attribute information for configuring the OWL is shown in Table 1 below. .

property summary rating Indicate your own preference for the event.
eventType Shows which event the corresponding post corresponds to. (Travel, watch, IT device reviews, etc.) subject Record the subject of the event. object Record the object of the event. beginTime Record the time when the event occurred. postingURI Record the address where you can access the post. reply Record the reply authors in the post. hasTitle Record the title of the post. hasText Record the body of the post.

The storage unit 330 includes an OWL file storage unit 332 in which an OWL file is stored as converted ontology based information, and a DBMS (DataBase Management System) 334 in which a body of a blog is stored.

The RDF store 400 stores an instance of the ontology of interest corresponding to the category, and preferably has a web communication function for communicating with the web-based blog server 300.

For example, the http protocol is supported using Sesame, an open source framework provided by the RDF store 400.

Hereinafter, a method of converting a blog post of the present invention into ontology-based information will be described with reference to FIGS. 3 to 7.

3 is a flowchart illustrating a method of converting a blog post into ontology-based information according to an embodiment of the present invention, and FIG. 4 is a diagram illustrating a web page before storage of a blog body (a) and a diagram showing a web page after storage ( b), FIG. 5 is a diagram illustrating an example of recommended attribute information for each tag, FIG. 6 is a diagram illustrating an example of the finally selected tag and its attribute information, and FIG. 7 is a diagram illustrating an example of the generated OWL file. .

In the method for converting blog posts of the present invention into ontology-based information, an input step (step S301) in which the user terminal 100 inputs a body of the blog and a category for the body, and the entity name of the inputted body into the category The tag recommendation step (step S302 to step S304) of searching in the corresponding ontology of interest and providing the user terminal 100 as a plurality of recommendation tags, and the user terminal 100 checks the plurality of recommended tags and selects one of them. A selection step of selecting (step S305) and a conversion step (steps S306 to S308) of converting and storing the input text into ontology-based information based on the selected recommendation tag.

More specifically, first, the user terminal 100 connects to the blog server 300 through the web browser 200 by the user and creates a desired blog post (step S301).

In this case, after the content is written, as shown in FIG. 4A, the user may select one of the predefined category categories so that the user can select a category related to the subject. Here, the category information received from the user serves to activate the ontology of interest.

When the user terminal 100 stores the text, as shown in FIG. 4B, the tag recommendation is activated to analyze the input text to recognize the entity name and extract the entity name as a tag candidate (step S302). .

That is, the entity name recognition unit 312 extracts entity names that can be tag candidates based on the text created by the user.

It is searched whether the extracted entity name exists in the ontology of interest (step S303).

For example, an instance of the ontology of interest stored in the RDF store 400 is searched to see if the entity name exists.

Next, the searched entity name is provided as the recommendation tag to the user terminal 100 together with the searched attribute (step S304).

In this case, it is preferable that the searched recommendation tag is compared with the word frequency and the attribute information and ranked to provide only the top N items. In addition, the provided attribute may include direct input.

Here, the recommended tag is an instance of the ontology of interest selected according to the category information.

As such, the recommended tag attribute changes according to the event information selected by the user. In other words, it is guided to annotate who, where, with whom, when, etc. for the event of 'go to travel'. Examples of attribute information of recommended tags according to an event category selected by a user are shown in Table 2 below.

category Tag attributes Remarks
IT
IT Product Name Review date (default: current date) grade (default: 3)
book
Book title Review date (default: current date) grade (default: 3)

Travel

Traveler's name Companion Travel period (default: current date) Transportation grade (default: 3)

In this case, as shown in FIGS. 5 and 6, since the recommended tag may have the same entity name having different attribute values, the user may be provided with attribute information about them to exclude ambiguity generated when selecting a tag. .

The user terminal 100 selects a desired tag by checking the provided recommendation tag list (step S305).

That is, when the tag list existing as an instance of the ontology of interest is provided to the user terminal 100, the user selects a desired tag suitable for the written text.

When the user selects a desired tag, the input text is converted into ontology based information according to the selected tag and its attributes (step S306).

In this case, the user terminal 100 may select a direct input, which is one of the attributes, and input a desired tag.

On the other hand, if the user terminal 100 does not select the recommendation tag, the input body is converted into ontology-based information according to the default recommendation tag and its attributes (step S307).

Here, it is preferable that the ontology-based information is composed of an OWL file.

For example, based on the structure of the OWL file constructed using the information shown in Table 1, an OWL file as shown in FIG. 7 is generated using contents corresponding to each post. The RDF language is used to create these OWL files.

At this time, it is set to a file name having the same value as the ID of the corresponding post so that it can be easily collected at a later search and stored in the blog server (step S308).

Through this recommendation process, the user can easily attach a tag with semantic information, and the stored post is generated as an OWL file and collected in the blog server 300.

In this way, more accurate blog ontology can be built.

The present invention described above is not limited to the above-described embodiments and the accompanying drawings, and it is common in the art that various substitutions, modifications, and changes can be made without departing from the technical spirit of the present invention. It will be apparent to those skilled in the art.

1 is a block diagram schematically illustrating a system for converting a blog post into ontology-based information according to an embodiment of the present invention.

2 is a block diagram showing a detailed configuration of the blog system of FIG.

3 is a flowchart illustrating a method of converting a blog post into ontology-based information according to an embodiment of the present invention.

4 is a view (a) showing the web page before the storage of the blog body and (b) shows a web page after the storage,

5 is a diagram illustrating an example of recommended attribute information for each tag.

6 is a diagram illustrating an example of a finally selected tag and its attribute information;

7 is a diagram illustrating an example of a generated OWL file.

Explanation of symbols on the main parts of the drawings

10: conversion system 100: user terminal

200: Web Browser 300: Blog Server

310: tag attachment portion 312: object name recognition portion

314: tag recommendation unit 316: search unit

320: conversion unit 330: storage unit

332: OWL file storage unit 334: DBMS

400: RDF Store

Claims (15)

An input step of the user terminal inputting a body of the blog and a category for the body through a web browser; A candidate extraction step of analyzing the inputted text to recognize the entity name and extracting the entity name as a tag candidate, and searching for whether the extracted entity name exists in an instance of the ontology of interest corresponding to the category; And a tag recommendation step of comparing and ranking the retrieved entity names with word frequency and attribute information and providing them with the attributes as the top N recommended tags. A selection step of checking, by the user terminal, the recommended plurality of tags and selecting one of the tags; A conversion step of converting the input text into ontology based information based on the selected recommendation tag and converting the input text into ontology based information according to a default recommendation tag and its attributes when there is no selection of the user terminal; How to convert a blog post containing the ontology-based information. delete delete The method of claim 1, The providing step converts the blog post into ontology based information, characterized in that the direct input is provided as an attribute. delete The method of claim 1, And the converting step converts the input text into ontology-based information according to the selected tag and its attributes. delete The method of claim 1, The converting step is a method for converting a blog post to ontology-based information, characterized in that the ontology-based information is composed of Ontology Web Language (OWL) file. A user terminal accessing a blog through a web browser and inputting a text and a category thereof; A resource description framework (RDF) store in which an instance of the ontology of interest corresponding to the category is stored; A tag attaching unit which searches for the entity name of the input text in the RDF store and provides the recommendation tag to the user terminal as a plurality of recommendation tags, and converts the input text into ontology-based information based on the recommendation tag selected by the user terminal. Including; Blog server including a conversion unit to include, The tag attaching unit, the tag attaching unit analyzes the input text and extracts an entity name, and compares the recommendation tag with the word frequency and attribute information and ranks the recommendation tag, and provides only the top N as a recommendation tag, A tag recommending unit providing the recommendation tag selected by the user terminal to the converting unit, and a searching unit searching whether the extracted entity name is an instance of the ontology of interest of the RDF store, and searching the N recommendation tags in the RDF store. Providing to the user terminal along with an attribute, providing the recommendation tag selected by the user terminal and its attributes to the transformation unit, and providing a default recommendation tag and its attributes to the transformation unit if there is no selection of the user terminal. A system for converting blog posts into ontology-based information. The method of claim 9, And the conversion unit converts the input text into ontology-based information according to the selected recommendation tag and its attributes, and the default tag and one of the attributes. The method of claim 9, The blog server further comprises a storage unit for storing the converted ontology-based information blog post to ontology-based information conversion system. The method of claim 11, And the conversion unit converts the blog post into ontology-based information, wherein the ontology-based information is generated as an OWL file. 13. The method of claim 12, The storage unit, An OWL file storage unit in which the OWL file is stored; System for converting blog posts into ontology-based information, characterized in that it comprises a DBMS (DataBase Management System) that stores the body of the blog. The method of claim 9, The RDF store has a web communication function, the blog post to ontology-based information conversion system. The method of claim 9, The recommendation tag is a blog post to ontology-based information conversion system, characterized in that the direct input is included as an attribute.

Images