KR20050000468A - A Method For Classifying Document Information Based On User's Definition And Storage Media Thereof - Google Patents

Abstract

PURPOSE: A learning-based bibliographic information classifying method by a classification code and user classification definition, and a recording medium thereof are provided to classify bibliographic information efficiently by automatically classifying the bibliographic information if a user fixes a desired classification item/standard. CONSTITUTION: Classification object file selection information and classification code information are received from the user(S16). The bibliographic information included in a classification object file is classified by the classification standard information corresponding to the classification code and outputs a result to the user(S18). The selection information of the classified bibliographic information is received, and the corresponding abstract information and indexed keywords are extracted/transmitted to the user(S18). A keyword selection signal is received from the user. The selected keyword is added, edited, or removed from the classification standard information(S24).

Description

A method for classifying document information based on user's definition and storage media thereof}

The present invention relates to a method for classifying document information, and more particularly, classifies it as a specific classification code to provide a user with summary information and keyword information about the document information, and the user can easily classify a classification code corresponding to the classification code based on this. The present invention relates to a classification method that can quickly classify a document information by allowing it to be reinforced.

As the industry advances and the age of advanced technology is increasing, the importance of intellectual property rights, such as patents, is being emphasized day by day. Therefore, governments and corporations, etc., have patent information such as patent maps for the purpose of technology development and sales strategy. We are trying to grasp new technology trends through analysis of literature information, including.

In order to prepare such a patent map, a process of analyzing and classifying a myriad of patents in related fields is required. However, the conventional patent analysis and classification work takes a method of repeatedly checking by the intuition of the classifier while reading the abstract data by opening the patent data downloaded by the user one by one.

Therefore, this method may vary the quality of the classification data according to the quality of the classifier, there is a problem that excessive investment of time and effort because the data must be processed by reading each one.

In addition, since patent information is classified on a user's computer, it is very difficult to share the classification information with other classifiers, and the above process must be repeated to apply the same classification method to the same data group. It is true.

Accordingly, there is an increasing demand for a faster and more accurate document information classification method, overcoming the unreasonability of the document information classification method including the conventional patent information.

The present invention has been made in order to solve the above problems, the first object of the present invention is to set the classification items and classification criteria suitable for the user's purpose, such as patent documents, etc. The present invention provides a method of classifying document information based on a user classification definition and a recording medium for automatically classifying document information so as to quickly and efficiently classify document information.

The second object of the present invention is to define a user classification that can be reinforced by adding, deleting, and modifying classification criteria information while the classification results are displayed on the screen in a structure that can be easily understood by the user. It provides a method of classifying document information based on learning and a recording medium thereof.

A third object of the present invention is to provide a learning-based document information classification method based on a user classification definition in which a classification result is reflected in a classification standard and learning / evolution of the classification standard, and a recording medium thereof.

1 is a block diagram of a patent information classification system according to the present invention;

FIG. 2 is a flowchart illustrating a functional module of the server computer of FIG. 1 and data processing procedures thereof;

3 is a view showing only the automatic classification module among the functions of the server computer of FIG.

4 is a flowchart illustrating a process of classifying document information according to the present invention;

5 is a flowchart illustrating a process in which classification reference learning of document information is performed;

6 is an exemplary execution screen of a user computer on which a patent information classification method according to the present invention is performed.

<Brief description of the major reference numerals>

10: user computer ,,

20: server computer,

30: network.

According to a first aspect of the present invention for achieving the above object, in the learning-based document information classification method according to the user classification definition, receiving classification target file selection information and classification code selection information from a user computer, Classifying the document information included in the classification target file according to the classification criteria information corresponding to the selected classification code and transmitting the result to the user computer; selection information of the literature information classified as classification criteria information of the classification code Receiving and extracting corresponding green information and at least one or more indexed keywords to the user computer, receiving a keyword selection signal for selecting one or more keywords from the extracted keywords from the user computer, Define selected keywords from user's computer Performing the processing of any one of addition, modification, and deletion in the beam; and when verification request information about the changed classification criteria information is received from the user computer, the document information is reclassified based on the classification criteria information to reclassify the result. Transmitting to a user computer; And classifying the learning-based document information classification method according to the user classification definition, comprising: storing the result of the classification criteria.

According to a second aspect of the present invention for achieving the above object, a recording medium on which a computer-readable program is recorded, and a recording medium on which a computer-readable program is recorded, based on learning by user classification definition. A method of classifying document information of a document, the method comprising: receiving classification file selection information and classification code selection information from a user computer, classifying the document information included in the classification file according to classification criteria information corresponding to the selected classification code; Transmitting the result to the user computer, receiving selection information of document information classified as classification reference information of the classification code, extracting the corresponding green information and at least one indexed keyword, and transmitting the selected information to the user computer; Step, among the extracted keywords from the user computer Receiving a keyword selection signal for selecting one or more keywords, performing any one of the process of adding, modifying, or deleting the selected keyword from the user computer to the classification criteria information, for the changed classification criteria information from the user computer When the verification request information is received, the recording medium records the program for reclassifying the document information based on the classification criteria information, transmitting the reclassified results to the user computer, and storing the classification criteria results. Can be.

Other objects, specific advantages and novel features of the present invention will become more apparent from the following detailed description and the preferred embodiments associated with the accompanying drawings.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. The present invention relates to a method for classifying document information, which is possible to classify all document information available to the user in the form of output such as patent information, technical information, article information, news articles, historical feeds, etc. The classification method will be described by way of example. Of course, the patent information classification method disclosed below can be equally applied to other document information.

1 is a block diagram of a patent information classification system according to the present invention.

As shown in FIG. 1, the patent information classification system according to the present invention includes a user computer 10 and a server computer 20.

The user computer 10 is generally an IBM PC and is connected to the server computer 20 via a network 30 in a TCP / IP or PPP manner. Here, the network 30 includes the Internet, a LAN, a MAN, a WAN, and the like, and thus, may be implemented as a server / client structure based on a local network in addition to the web. The user computer 10 transmits data to be used for classification of patent information or a file to be classified to the server computer 20, and receives patent information classification result information and keyword information for indexing patent information from the server computer 20.

The server computer 20 is a computer installed in a company that provides patent information analysis and classification service. The server computer 20 classifies a plurality of patent information included in a classification target file received from the user computer 10 into a classification code set by a user. It performs the function.

FIG. 2 is a flowchart illustrating a functional module and a data processing procedure of the server computer 20 of FIG. 1.

Referring to Figure 2, the document processing module 50 of the server computer 20 is processed based on the processing procedure such as analysis / collection / input / protocol inspection / conversion based on the raw data 40 of the patent and literature Stored in the database 60. The processing module 50 processes the raw data 40 to include bibliographic matters such as the name of the invention, patent number, and the like, and text summary information summarizing the composition of the invention to facilitate a user's patent classification. Generate patent information.

The index module 80 indexes the patent information and the supplementary information collected by using the stopword dictionary 70 and the terminology dictionary 90 in accordance with the database table standard so that it can be applied to the search, analysis system, and automatic classification module. Upload to the database (110). There are two types of indexing methods: automatic indexes and manual indexes. The automatic index extracts the keywords processed by the processing module 50 through pre-reference / morphological analysis using the file information, and indexes the procedure of updating the indexed keywords in the index file / surgery / text file. In this way, indexing is done in batches without going through the work of the indexer. Manual indexing is a method in which the indexer indexes the processed data one record unit for each indexing procedure. The manual indexer extracts the terms directly from the data and the semi-automatic indexing system checks and corrects the indexes automatically. have. In this embodiment, an automatic indexing method is used.

The retrieval module 100 is a module for retrieving patent and document information by a user-specified search expression or natural language query. The search method is classified classification that accesses the classification group such as IPC classification and target classification according to the type and performs classification, term search using logical function of keywords, field search using index attribute by field, text by natural language query Search, citation search for specific patent data and citation data.

The automatic classification module 120 is a module for classifying the results retrieved by the search module 100 into a plurality of categories, that is, classification codes, according to a predetermined classification criterion. A process in which the automatic classification process is performed in the automatic classification module 120 will be described in detail later with reference to FIG. 3.

The project management module 130 integrates and manages the results created by the user in various ways such as a search, an analysis system, an invention support, and an automatic classification module in the form of a user file 140. This file is then shared with other users in the same workgroup. In order to allow other users to share the user file 140 corresponding to a specific user, each user can share only the users belonging to the same group as the user by using the user ID entered during the login process. The method may be applied such that only the other user designated by the user can share the same.

Although not shown in FIG. 2, an analysis module for analyzing a search result by the search module 100 may be further included. The conventional analysis tool has a limitation in that it is possible to download search data and analyze the data only on the terminal (for example, a computer) of the user. However, the analysis module according to the present invention can list the search results of data on the web using OLAP (On Line Application Processing) technology to analyze in real time. The analysis module according to the present invention may perform analysis by applying various conditions such as a patent number, an application date, an applicant, a title, an IPC code, an inventor, and the like. In addition, as an analysis result display method, a search result for each condition (for example, the number of applications for each applicant) may be compared and displayed in the form of a graph or the like. For example, the analysis module may include country analysis (i.e., extract the number of data corresponding to a registered country or filing country), technical analysis (i.e., extract the number of data by IPC code), and applicant analysis (extract the number of data corresponding to each applicant). , Year-by-year country analysis, and applicant country-specific analysis (i.e., extracting the data of applicants from each country).

3 is a diagram illustrating a configuration of a classification module of the server computer 20 of FIG. 1.

The classification items set by the user are stored in the classification item database 150, and the classification criteria for each classification item are stored in the classification reference database 160.

The patent list file 170 stores a list of patents that a user wants to classify.

The automatic classification program classifies the patent information corresponding to the patent list file 170 designated by the user with reference to the information stored in the classification item database 150, the classification criteria database 160, and the patent information database 200.

The classification result is displayed on the classification result screen 180, and the user repeats the process of reinforcing the classification criteria again after checking the classification result so that the classification criteria are gradually learned in a direction more in line with the classification intention of the user. .

The classification result may be stored as the classification result file 190 again.

When a plurality of classification result files 190 are generated by a plurality of users, the project management module may generate and manage one user file 140 by integrating the classification information files or the patent information classified by the plurality of users. It may be.

4 is a flowchart illustrating a process of classifying document information according to the present invention.

First, the patent information (target file), classification classification, classification code, etc. retrieved from the user computer 20 are selected and transmitted to the server computer 20 (S1).

The server computer 20 automatically classifies the patent data belonging to the classification criteria for each classification code based on the classification definition received from the user computer 20 with respect to the received patent data (target file) (S2).

The automatic classification process extracts the patent data belonging to the classification criteria for each classification code based on the classification definition (classification, code) received from the user computer 20 for the patent data (target file) received at the server computer 20. do. If the user does not specify a classification code or selects a whole classification, the patent information is classified by each classification code. Each patent data may correspond to several classification codes. Are classified. Here, the classification criteria for each classification code is a word or index word intended by the user or the words are combined using logical operation symbols (+, *). Each patent data is classified by each classification code through a corresponding logical operation process.

When the server computer 20 transmits the patent information of which the automatic classification is completed to the user computer 20 (S3), the automatically classified patent information is displayed in a tree structure on the screen of the user computer 20 (S4). That is, the server computer 20 displays the number of patents for each classification code of the patent data for which the automatic classification is completed in the form of a directory, and if each classification code node is selected, a list of the corresponding patent data is displayed as a lower node. When the specific patent data of the lower node is selected and transmitted to the server computer 20 (S5), the server computer 20 extracts detailed information (book matters and body summary information) about the patent data (S6). .

When the server computer 20 transmits the extracted patent details (book details and body summary information) to the user computer 20 (S7), the bibliographic details and body summary information of the patent are displayed on the web screen of the user computer 20. Is displayed (S8). In this embodiment, multiple views are supported only when the number of automatically categorized patents is 20 or less in consideration of the delay of the loading speed.

The user computer 20 transmits the keyword extraction signal to the server computer 20 (S09). That is, when a specific code of the patent abstract information displayed on the screen of the user computer 20 is selected, the server computer 20 sends the corresponding patent information. Extract the keyword corresponding to and transmit it to the user computer 20 (S10, S11). Here, keywords are extracted by morphological analysis of the words included in the text summary of the patent information, removing unnecessary words and investigations, and calculating the frequency and weight of each noun to extract proper nouns. However, the present invention is not limited thereto, and there may be various methods of extracting keywords.

The process of displaying and reviewing the results of the automatic classification on a screen and reinforcing the rules of the classification criteria will be described in FIG. 3.

This automatic classification process can be repeatedly executed for the patent information retrieved for a specific purpose (S12).

When the user computer 20 transmits the "save" request signal for the patent information for which the automatic classification is completed to the server computer 20 (S13), the server computer 20 stores the automatically classified patent information data in the form of a file in a predetermined form. After the storage (S14), when the storage is completed, the contents of the storage is displayed on the user's computer (S15).

5 is a flowchart illustrating a process of performing classification criterion learning of document information.

Referring to FIG. 5, first, a target file, classification criteria, and classification code are selected by the user computer 20 and transmitted to the server computer 20 (S16).

In the present invention, the classification code and the classification code may be provided in addition to the basic information, and various methods may be introduced, including a user can directly write the classification.

The server computer 20 extracts classification criterion information corresponding to the received classification classification and classification code (S17) and transmits it to the user computer 20 (S18).

The user selects a key keyword in defining the classification criteria for each classification code (S19). In selecting a key keyword, the user selects any keyword appropriate for the classification code, or selects an index keyword from a text summary of the patent data.

The user creates an optimal classification standard by registering new classification criteria or reinforcing existing classification criteria using the selected key keywords (S20). Here, you can add, modify, or delete using the logical operators * (and) + (or) or parentheses around keywords.

The user executes a process of confirming (testing) whether the enhanced classification standard information is well prepared. If a test is desired, the user computer 20 transmits the classification standard information, the classification code, and the classification target patent information to the server computer 20. It transmits (S21). Here, the server computer 20 applies the information received from the user computer 20 to automatically classify the patent information data of the target file (S22), and displays the result of the automatic classification on the screen of the user computer 20 (S23). ). In addition, if the storage classification criteria information needs correction, correction, deletion, etc., the classification learning process is repeated from the selection of new keywords.

The user checks the result of reclassification automatically and continuously reinforces the dispute standard information until a satisfactory result is displayed. If the user reviews the displayed reclassification result and the result is satisfied, the user adopts the classification criteria information (S24).

The server computer 20 stores the classification criterion information received from the user computer 20 (S25), and displays the storage result on the screen of the user computer 20 (S26).

6 is an exemplary execution screen of a user computer on which a patent information classification method according to the present invention is performed.

When the user connects to the server computer 20 and selects the patent information automatic classification menu, a screen as shown in FIG. 6 is displayed.

The target file window 200 is a window for designating a file in which patent information to be classified is stored. That is, when the user clicks on the target file window 200, a list of target files related to the user is displayed in a pull-down manner, and the user selects a classification target file from the displayed list.

The classification partition window 210 is a part for displaying a classification classification registered by an IOC (functional classification), ICC (technical classification) or a user.

The classification code window 220 is dynamically generated according to the selected classification classification, and the method classification and the function classification are displayed according to the classification classification.

After the user selects the target file, classification classification and classification code, and clicks classification start 230, classification processing of patent information is performed.

The classification tree window 240 displays a classified result in a folder tree form according to the classification code. One patent information may belong to several classification codes, and patent information not belonging to any classification code is classified as an unclassified patent. 5 shows a case where the user selects the entire classification as the classification code. Accordingly, 20 classification codes are displayed in the classification tree window 240. If the user clicks on any one of the 20 classification codes, the patent information classified by the classification code is displayed in a hierarchical structure. FIG. 5 shows that six patent information is displayed when a classification code of "dividing" is clicked.

The abstract output window 250 is a portion in which an abstract including bibliographic matters and text summary information of patent information classified under a designated classification code is displayed. As described above, 20 patent information belonging to a classification code designated as a classification result is displayed. If less, the patent information is output in multiple windows so that the user can easily search for and review the patent information. However, this is in consideration of the loading speed, it is also possible to support a multi-view function for more than 20 patent information, if necessary. If the user clicks on any one of the six patent information, as shown in Fig. 5, only the abstract of the selected patent information is displayed.

The index word window 270 is a portion for displaying an index word corresponding to the selected patent information, and the index word refers to a keyword extracted by the index module 80. As described above, keywords may be extracted through techniques such as morphological analysis and frequency analysis.

The classification criteria keyword window 260 is a portion in which classification criteria keywords for classifying patent information into specific classification codes are displayed. If a user clicks on a specific keyword in the index word window 270, the keyword is displayed in the classification criteria keyword window 260. In this case, when selecting a plurality of keywords, the keywords must be combined through the operator using the operator button 280, and if there is a continuous input of the keyword without the operator, a message for requesting the operator input is output.

After the user adds the keyword, click the registration button 300 to register a new classification criteria. If the same contents are re-registered, a message indicating that they are already registered is displayed.

When the verification button 310 is clicked, the reclassification is performed on the classification code selected as the reinforcement classification standard by adding, deleting, or modifying in the window 260 in which the classification standard is displayed for the patent information selected in the target file window 200. After executing, the result is printed out as a message.

After checking the verification result, the user clicks the apply button 320 to apply the enhanced classification criteria to the classification criteria database of the server computer 20.

If the user clicks on the storage result of classification according to whether the user is satisfied with the classified result, the result classified by the classification criteria is stored in the form of a schedule file in the project management, and the result is stored as a message.

Although the exemplary embodiment has been illustrated and described with reference to the method of classifying patent information, the present invention is not limited thereto, and the present invention can be applied to a classifying method of various types of information.

According to the present invention, there is an effect that the user can set the classification definition of the document information (for example, the form of the output, such as patent information, technical information, article information, news articles, historical feeds, etc.) suitable for his purpose. .

In addition, the classification of documents can be performed quickly and efficiently, and the classified results can be easily and effectively grasped.

In addition, the classification result is reflected in the classification criteria again, and thus the classification criteria are learned / evolved.

Although the present invention has been described in connection with the above-mentioned preferred embodiments, it is possible to make various modifications or variations without departing from the spirit and scope of the invention. Accordingly, the appended claims will cover such modifications and variations as fall within the spirit of the invention.

Claims (13)

In the learning-based document information classification method by user classification definition, Receiving classification target file selection information and classification code selection information from a user computer; Classifying the document information included in the classification target file according to classification criteria information corresponding to the selected classification code and transmitting the result to the user computer; Receiving selection information of document information classified as classification reference information of the classification code, extracting corresponding green information and at least one or more indexed keywords, and transmitting them to the user computer; Receiving a keyword selection signal for selecting one or more keywords among the extracted keywords from the user computer; Performing any one of a process of adding, modifying, or deleting a keyword selected from the user computer to classification criteria information; Reclassifying the document information based on the classification criteria information and transmitting the reclassified result to the user computer when the verification request information on the changed classification criteria information is received from the user computer; And And storing the result of the classification criteria. The method of claim 1, Receiving a reclassification request signal from the user computer; Reclassifying the document information; Receiving a classification result storage request signal from the user computer when the reclassified document information matches the classification criteria information defined by the user computer; And And classifying the classification result into a user classification result file. The method of claim 1, Receiving classification criteria information corresponding to a plurality of classification codes; And classifying the plurality of pieces of document information into the plurality of pieces of classification codes based on the classification criteria information. The method of claim 3, wherein Processing and storing raw document information; And And extracting and storing at least one or more index words from the processed document information according to a preset method. The method of claim 4, wherein The index word extraction is a learning-based document information classification method according to a user classification definition, characterized in that extracting only the root from the words included in the document information in a predetermined manner. The method of claim 5, wherein The index word is a learning-based document information classification method according to a user classification definition, characterized in that included in the document information more than a preset number. The method of claim 1, And configuring the screen so that all the document information classified by the classification criteria is displayed on one screen. The method of claim 7, wherein The bibliographic information is a patent number, IPC code, application date, publication date, name, applicant, inventor or summary information characterized in that the learning-based literature information classification method according to the user classification. The method of claim 7, wherein The method of classifying document information based on a user classification definition, comprising the step of configuring the document information classified by the classification criteria into the number of documents included in each classification code and the corresponding document by patent number and name. . The method of claim 1, And storing the reclassified document information as shared information of a specific project group. The method of claim 10, Receiving the classified or reclassified document information request signal including user information from another user computer belonging to the specific project group; Confirming whether the other user is a member of the specific project group based on the user information; And And when the member authentication is completed, transmitting the classified or reclassified document information to the other user computer. The method of claim 11, And generating an integrated classification target file by integrating the document information classified by the respective user computers belonging to the specific project group. A recording medium having a computer readable program recorded thereon, In the learning-based document information classification method by user classification definition, Receiving classification target file selection information and classification code selection information from a user computer; Classifying the document information included in the classification target file according to classification criteria information corresponding to the selected classification code and transmitting the result to the user computer; Receiving selection information of document information classified as classification reference information of the classification code, extracting corresponding green information and at least one or more indexed keywords, and transmitting them to the user computer; Receiving a keyword selection signal for selecting one or more keywords among the extracted keywords from the user computer; Performing any one of a process of adding, modifying, or deleting a keyword selected from the user computer to classification criteria information; Reclassifying the document information based on the classification criteria information and transmitting the reclassified result to the user computer when the verification request information on the changed classification criteria information is received from the user computer; And A recording medium having recorded thereon a program for executing the step of storing the result of the classification criteria.