KR101401225B1 - System for analyzing documents - Google Patents

Abstract

The document analysis system of this embodiment includes a database in which patent documents are stored; A document classifying module for reading an indirect citation relation between the patent documents and performing classification on patent documents to be analyzed by using the indirect citation relation read out; A document evaluation module for performing an evaluation on the patent document and calculating an evaluation value as a result of the evaluation using the attribute information of the patent document; And UI output means for providing the user with evaluation information on the patent documents to be analyzed based on the attribute information.

Description

System for analyzing documents

The present invention discloses a system for performing clustering and automatic classification of a plurality of patent documents using an indirect citation relationship between documents, and for performing analysis and evaluation on documents for which classification has been performed.

When a patent applicant intends to obtain a patent, it is necessary to prepare a document that meets the prescribed requirements and submit it to the Patent Office. The patent application documents submitted to the Patent Office are released after a predetermined period of time or requirements have been met, and these documents can be called patent documents.

Generally, a person who intends to apply for a patent is subjected to a process of searching / searching for such patent documents in order to confirm the existence of prior art. Most patent document searches are performed by inputting keywords.

That is, with the recent development of technology, the number of patent applications has become enormous, and accompanying this, the amount of patent documents has become enormous. Accordingly, it is not easy to prevent duplicate research, to confirm whether or not infringement has occurred, to conduct prior art research before patent application, to grasp trends of technology development of other companies, and to investigate patent documents for improvement of research and development.

In a conventional search system for searching or examining such a patent document, if the keyword is selected incorrectly, there is a case where unnecessary information is included in a large amount. In such a case, there is a problem that the time of the investigation itself becomes enormous.

Embodiments of the present invention can perform more efficient classification and clustering between patent documents by deriving references or quotation relationships among a plurality of patent documents or by reading indirectly quoted relationships even if they are not directly related to each other, And to provide a branch analysis system that can more effectively provide the user with the analysis results.

The document analysis system of this embodiment includes a database in which patent documents are stored; A document classifying module for reading an indirect citation relation between the patent documents and performing classification on patent documents to be analyzed by using the indirect citation relation read out; A document evaluation module for performing an evaluation on the patent document and calculating an evaluation value as a result of the evaluation using the attribute information of the patent document; And UI output means for providing the user with evaluation information on the patent documents to be analyzed based on the attribute information.

The attribute information includes external characteristics derived by taking into account the internal characteristics derived from the matters recorded in the patent document and the matters recorded in the patent document for citing the patent document.

The attribute information further includes a quotation index that uses information on the number of citations or the number of citations between the patent documents.

The document evaluation module classifies the evaluation result of the patent document into a third party patent other than the pre-established patent, and the UI output means outputs, as the evaluation information, the first evaluation information And second evaluation information about other patents.

According to the embodiment as proposed, there is an advantage that a more effective classification between patent documents can be performed by deriving a reference or quotation relation between a plurality of patent documents, or by reading an indirect quotation relation even if the quotation is not directly related.

In addition, there is an advantage that a user can easily analyze a patent document by providing information to a user through various UIs as a result of classification and clustering of an effective document.

1 is a diagram for explaining a document analysis system according to the present embodiment;
2 is a diagram for explaining the operational flow of the document analysis system according to the present embodiment;
3 is an example of an evaluation factor table of the document evaluation module according to the present embodiment.
4 is an example showing the result of search and evaluation of a document according to the present embodiment.
5 is an example of a UI in which information of a document is displayed according to the present embodiment.
6 is a diagram showing a configuration of document grouping means according to the present embodiment;
FIG. 7 is a diagram for explaining an indirect citation relationship according to the present embodiment; FIG.
8 is a diagram for explaining how a second group document is classified and grouped into categories of a first group according to the present embodiment;
9 is an example showing attribute information of the category document or the second group document according to the present embodiment.
10 is an example showing a feature vector calculated from a category document or a second group document according to the present embodiment.
11 to 15 illustrate various types of UIs provided to a user as a result of document classification and clustering according to the present embodiment.

Hereinafter, the present embodiment will be described in detail with reference to the accompanying drawings. It should be understood, however, that the scope of the inventive concept of the present embodiment can be determined from the matters disclosed in the present embodiment, and the spirit of the present invention possessed by the present embodiment is not limited to the embodiments in which addition, Variations.

In the following description, the word " comprising " does not exclude the presence of other elements or steps than those listed.

1 is an example showing a configuration of a document analysis system according to the present embodiment.

A system such as the one shown in FIG. 1 may include a database 130 in which a patent document is stored, and a predetermined evaluation factor (user) for a patent document stored in the database 130 or other patent documents accessible via a network A document evaluating module 140 for evaluating the patent documents by using the patent documents stored in the database, and direct and indirect citation relations of the patent documents designated by the user or stored in the database are derived so that classification and clustering of the patent documents are performed And a document classification module 150. [

In addition, such a document analysis system classifies the patent documents using the indirect citation relation between the patent documents, and uses the result of the classification (the representative document in the classified patent documents) Output module 110, the document retrieval module 120, the document characteristic creation module 160, and the document characteristic DB 170, which can be realized by a server device, a computer, or the like, .

2, the document analysis system includes a patent search step S101, a patent document analysis step S102, a document classification step S103 for analyzing patent documents, a UI (Step S104), and each step will be described in more detail using each configuration of the document analysis system.

First, the operation of searching for a patent document (S101) using the document analysis system will be described.

The query input unit 111 of the input / output module 110 receives a query input by a user using a keyboard, a mouse, or the like in order to perform an operation such as document search or analysis. 130) or a patent document stored in a patent document (which can be network-connected). In addition, the keyword includes numbers such as an application number, a public number, etc. constituting the patent document in addition to characters.

The user interface output unit 112 of the input and output module 110 outputs the information calculated or extracted by the document search module 120 or the document classification module 150 or the document evaluation module 140 to the user side And is provided as an apparatus for providing various UIs to be described later, but it is also possible to arrange it in other components of the evaluation system according to the embodiment.

In addition, the database 130 of the embodiment stores patent document data, and the patent document data group is a database configured to store document data of a specification relating to a patent application or patent to be electronicized. This patent document data is data including text data describing the content of the specification by the character code. Document data including a description in general purpose tag language such as SGML (Standard Generalized Markup Language), HTML (HyperText Markup Language), and XML (eXtensible Markup Language) elsewhere in plain text data is also possible. Other formats such as Portable Document Format (PDF), document format of a general word processor, and Rich Text Format (RTF) format are also possible if text data can be extracted.

The patent document database 130 may be provided outside the patent document evaluation system. In this case, the patent document evaluation system may access the database using the network, and acquire document data of the patent document.

The document retrieval module 120 retrieves patent documents to be retrieved from the patent documents stored in the database 130 based on the query input by the user. In searching for a patent document by the document search module 120, the document feature creation module 160 and the document feature DB 170 may be used.

The document characteristic creation module 160 may obtain texts from the documents stored in the database 130 and provide index information on the frequency of each keyword to the document characteristic DB 170. [ When a predetermined query term is received by the query receiving means 111, the document searching module 120 searches for documents including a query term using an index file of each document stored in the document characteristic DB 170 can do.

A document as a result of the search by the document search module 120 may be provided to the user through a UI output means 112 as shown in FIG.

When a predetermined query is received through the query receiving unit 111 or when a new document is stored in the database 130 by the web robot, the document characteristic creation module 160 creates an index file And the feature vector for each document can be determined using the generated feature vectors. This will be described with reference to FIG.

FIG. 9 is a diagram showing attribute information of each document. Attribute information of the documents shown in FIG. 9 can be created in the form of an index file by the document characteristic creation module 160, (170).

Using the index file stored in the document feature DB 170, the document feature creation module 160 can determine a feature vector of each document, and the feature vector can also be stored in the document feature DB 170.

FIG. 9 shows information on the frequency of occurrences of the keywords A, B, C, D, M, I, K, O, P, Q and Z for each document. For example, A is 35, the keyword B is 19, the keyword C is 15, the keyword D is 13, the keyword A is not Is included.

The occurrence frequency table for each keyword included in each document can be created so as to be sequentially arranged from the keyword having the highest frequency to the keyword having the lower frequency as shown in FIG.

For example, in Document 1, in order to indicate that the keyword A includes 4.5%, the keyword B contains 2.4%, the keyword C contains 1.9%, and the keyword D contains 1.7%, the index file for document 1 includes (A, B, C, D) → (4.5%, 2.4%, 1.9%, 1.7%).

An index file of each document is created by the various methods described above, and the feature vector of each document can be extracted by using the created index file.

In detail, the document characteristic creation module 160 creates a table based on the frequency of occurrence of each keyword in each document, and creates a feature vector of each document using the generated table.

Herein, the feature vector determined by the document characteristic creation module 160 has the evaluation value of the keyword as an element for each document. For example, when the total number of keywords included in each document is n, the feature vector Is a vector of n-dimensional space and can be expressed as the following equation (1).

Characteristic vector = (evaluation value w1 of keyword A, evaluation value w2 of keyword B, evaluation value wn of word n) --- (1)

For calculation of evaluation values, for example, the tf · idf method disclosed in Salton, G: Automatic Transformation, Analysis, and Retrieval of Information by Computer, Addison-Wesley can be used. According to the tf.idf method, a value other than 0 is calculated as an evaluation value for an element corresponding to a keyword included in the first document among n-dimensional feature vectors corresponding to the first document, and is included in the first document 0 " is calculated as an evaluation value for an element corresponding to a keyword (a word with a frequency of 0) which is not included in the keyword.

From this viewpoint, the evaluation value of the keyword as an element of the feature vector can be the frequency rate of each keyword appearing in each document. For example, by the document search module 120, a keyword A, a keyword B, and a keyword C from a first document can be clustered into a similar word, and the clustered similar words can be stored in a separate similar DB.

That is, the predetermined keyword A and keyword B are clustered by the document search module 120, and the clustered keyword A and keyword B are stored in the similarity DB.

If the extracted keyword includes any one of the keyword A and the keyword B, the document search module 120 searches for a similar document including the remaining keywords.

The retrieval of similar documents can be performed based on the attributes of the patent documents rather than the search limited to the extracted keywords.

When the keyword A is included in the query term received through the query receiving means 111, the search for a document including the keyword B and the keyword C together with the keyword A can be performed at the time of retrieving the similar document.

On the other hand, the document analysis system grasps the importance or trend of each document by assigning evaluation values to documents using a predetermined evaluation factor in a patent search process or a document classification and clustering process.

The evaluation of the patent document by the document analysis system will be described.

For the patent documents stored in the database 130 or accessible via the network, the document evaluation module 140 assigns evaluation values according to a predetermined evaluation factor.

That is, the document evaluation module 140 evaluates the patent document using the attribute information of the patent document with respect to the patent document stored in the database 130 or network connectable patent documents, And provides the result to the UI output means 112 to be displayed to the user. The UI output means 112 may provide the user with information about the evaluated value of the patent documents searched together with the search result list of the patent document, and may be provided with a pop-up or OSD It may also provide evaluation information on patent documents.

The document evaluation module 140 creates an evaluation item table using the evaluation items set for the patent documents stored in the database 130 or the network documents that can be connected to the network, And may be performed whenever a new patent document is stored in the database 130. [

However, the evaluation of the patent document by the document evaluation module 140 may be performed when there is a document search request of a user and a document to be searched exists. In the following description, Describe without limitation.

The document evaluation module 140 includes an evaluation factor management unit 141 that manages the characteristics of the patent document as an evaluation factor, a document management unit 141 that evaluates the patent document stored in the database 130 using the evaluation factor, Evaluation means 142 and DB document management means 143 for making the evaluation value, which is a document evaluation result by the document evaluation means 142, correspond to each of the patent documents.

The evaluation factor management unit 141 manages the internal and external characteristics of the patent document stored in the database 130, and these characteristics can be edited by the user.

That is, the structure of the evaluation factors for the internal and external characteristics of the patent document by the evaluation factor managing means 141 is shown in FIG. 3 is a diagram showing a structure of an evaluation factor of a patent document.

As shown in FIG. 3, a plurality of tables of the attributes relating to the patent described by the evaluation factor management means 141 can be linked for each country. Each table has internal characteristics derived from items recorded in the patent document, , And external characteristics that can be derived by considering the characteristics of the cited document cited by the patent document.

The internal characteristics that can be derived from the matters recorded in the patent document refer to the information of the keyword or the patent document that can be extracted through text mining operation on the description of the patent document.

For example, the maintenance period in which the period from the registration date registered to the current date recorded in the patent document is calculated can be derived from the matters described in the patent document, and thus can be an internal characteristic of the patent document.

The length of the claims, which can be determined according to the elapsed time information of the period from the filing date to the current date described in the patent document, the number of independent terms of the patent document, the number of keywords read out as a text mining result for a specific independent term, The number of subordinate terms that can be identified as subordinate terms can also be an internal characteristic of a patent document, since certain phrases such as 'in' or 'according to claim 1' are entered.

In addition, the number of inventors described in a patent document can also be an internal characteristic of a patent document.

However, the number of patents filed by the inventor A, which is recorded as the inventor in the first patent document, is the external characteristic of the patent document since the inventor A must retrieve other patent documents recorded as inventors.

When there are other patent documents cited in the patent document, the number of patent documents cited and the period for citation / citation are the external characteristics of the patent document.

In order to evaluate an evaluation value for scoring a patent document, an evaluation factor for a patent document must be defined and a weighting value for each of the evaluation factors defined is calculated to finally calculate an evaluation value for the patent .

In this regard, the evaluation factor management unit 141 creates evaluation factor items for each of the patent documents stored in the database 130, using an exemplary table as shown in FIG. FIG. 2 shows an example in which the internal characteristic and the external characteristic are randomly arranged, but the evaluation value of the internal characteristic obtainable from the information extracted in the patent document and the patent document and the other patent document And other patent documents in the same technical field that are stored in the database are also possible) may be distinguished as separate items.

After the values of the characteristics read out from the respective patent documents are recorded in the table as shown in Fig. 3, the evaluation value of the patent document is calculated by the document evaluating means 142. Fig.

For example, a predetermined weight may be assigned to each evaluation factor. In this case, the weight is calculated on the values of the internal characteristic and the external characteristic extracted from the patent document, so that the sum of the scores of the evaluation factors is It can be an evaluation value of a patent document.

Evaluation values of the patent document thus calculated can be separately managed by the DB document management means 143. When the information of the patent document search result is displayed to the user, evaluation values calculated for each patent document included in the retrieved result .

Therefore, the UI output means 112 of the input / output module 110 may provide the user with an item or table of evaluation factors managed by the evaluation factor management means 141, and may be provided with an evaluation factor Is stored and managed by the evaluation factor management means 141. [

Each of the patent documents can be given an evaluation value by the document evaluation module 140 as described above. When the patent document is displayed on the user side as a result of the search, Likewise, it can be seen.

For reference, FIG. 4 illustrates a list of document search results provided to a user's computer or server. For example, when seven search results of the patent document stored in the database 130 are read by the document search module 120 for the query term input by the user, the bibliographic information about the patent documents to be searched (For example, the patent number, the status, the filing date, the patent date, the name of the invention, the IPC) and the evaluation value of each patent document are displayed together.

In addition, the document evaluation unit 142 provides the UI output unit 112 with an evaluation value for the patent document so that the user can quickly distinguish the patent having the highest value from the patent documents having the retrieved result . In addition to the evaluation values for each of the patent documents, an average evaluation value of the retrieved patent documents may be calculated, and the average evaluation value may also be provided to the UI output means 112.

When the average evaluation value of the patent documents is searched together, the user can easily determine the superiority of each of the patent documents of the search result. According to the present embodiment, The search efficiency can be improved.

The document evaluation means 142 can calculate an average evaluation value in the technical field including the patent documents of the search result. The UI output means 112 outputs the evaluation value of each patent document of the search result The average evaluation value in the technical field to which the patent documents belong can be provided together.

In this case, the commonness of the technical fields to which the retrieved patent documents belong may be determined by the IPC classification of the international classification, or by the F-term classified by the Japanese Patent Office. When patent documents classified in different technical fields are to be output as search results, an average value of evaluation values for technology fields to which patent documents occupying a large proportion in search results belong can be provided.

Therefore, the user can easily grasp the degree of importance of the patent documents of the searched result by comparing the evaluation value given to each of the searched patent documents with the average evaluation value of the patent documents of the technical field .

On the other hand, when the search result list is downloaded, the information on the evaluation value performed by the document evaluation module 140 may be provided to the user's computer or To be provided to the server.

In the UI of the search result as shown in Fig. 4, when the user clicks a specific evaluation value (Weighting Value) in order to check the details of the evaluation value given to each patent document, And a separate UI for allowing the detailed evaluation of the score given to the patent document for each evaluation factor can be provided.

In addition, in a UI including a list of search results as shown in FIG. 4, when a user selects a specific patent document, a separate window (UI) showing the summary contents of the patent document may be created have. That is, as shown in FIG. 5, a patent document analysis UI can be provided to the user side, and evaluation information on the corresponding patent document is also provided to the UI.

For example, an item of an evaluation factor applied to the patent document, and information of a score of each item may be provided, along with a description of the invention, a representative drawing, and a summary of the selected patent document. And, as described above, the average evaluation factor values of the retrieved patent documents or patent documents of the same technical field as the patent may be provided together.

Then, the user can manipulate his / her server, computer, etc. to modify and edit the displayed evaluation factor item, and the assigned score can also be edited separately. To this end, the evaluation factor management unit 141 and the DB document management unit 143 of the document evaluation module 140 change the information of the corresponding patent document so as to correspond to the item and score of the evaluation factor changed by the user.

Next, after the process of evaluating the documents stored in the database or performing the evaluation work on the separate documents requested by the user, the document classification module 150 of the present embodiment (S103) is performed. As a result of this operation, information of various UIs as shown in FIG. 11 to FIG. 15 can be provided to the user by the UI output means 112. FIG.

The document retrieval module 120, the document evaluation module 140, and the document classification module 150 of the present embodiment may be used together with a predetermined algorithm in order to make retrieval, classification, and clustering of documents more effective, It can be seen that it operates in a complex manner.

Hereinafter, when predetermined patent documents are retrieved by the document retrieval module 120 and the document characteristic creation module 160 with respect to a query input by the user, and the retrieval result is shown in the list as shown in FIG. 3, An operation of classifying documents that are similar to the technical solution (problems in the prior art) or the solution (the solution to the problem) to the patent documents in the document search result will be described.

That is, according to the present embodiment, documents can be classified by using the indirect citation relation between patent documents, and since patent documents having such a citation relation tend to share the technical solution or solution, the database 130 It is more advantageous to classify the patent documents of the result of document search (including similar search) of the query term inputted by the user, rather than classifying all of the patent documents stored in the document database.

In this regard, the operation of the document classifying module 150 will be described by exemplifying patent documents belonging to a predetermined similar range as a result of document retrieval. However, the document evaluation module 140 operates in clustering of documents after classification of patent documents, but evaluation value information given as shown in FIG. 4 can also be provided in the document classifying step and the document searching step before document clustering.

Meanwhile, the UI output means 112 includes a tag 34 (see FIG. 4) for guiding the user to perform classification and clustering on some patent documents or all patent documents of the search results from the list of patent documents of the search result, Can be provided.

When the key for requesting document classification and clustering is input, the document classification module 150 derives an indirect citation relation for selected patents and performs document classification using the indirect citation relation. For example, when the first patent document is cited in the second patent document and the second patent document is cited in the third patent document, since the first patent document and the third patent document are indirectly quoted, The document classification module 150 classifies the first and third patent documents into the same category together with the second patent document.

Here, the citation relation according to the present embodiment, that is, the indirect citation relation will be described. Quotation relationship means that, if the reference document number (patent application number, patent publication number, registration number, etc.) of another patent document described in the patent document to describe a problem of the prior art is described, And the patent document to be cited can be formed.

In addition, the patent document referred to or described in the patent document need not be limited to the document to be cited, and the document referenced by the prior art / cited invention in the examination, objection or invalidation judgment of the patent document It can be said that it is in the citation relation. Accordingly, not only the case where the bibliographic information of other patent documents are described in the patent document but also other patent documents which can be used indirectly by the examiner or other third party during the examination can be said to be in the citation relation .

In order to expand the citation relation, the database 130 may be provided with a citation and a reference document storage section for storing information on whether or not each patent document is cited. In this case, A reading means for deriving a citation relation from the documents used during the examination or the post-registration procedure may be provided from the materials provided by the Patent Office and the like.

For example, if the disclosure number of another B patent document is listed in the A patent document, the direct quotation relation between the A patent document and the B patent document can be read. In addition, if the C patent document is presented by the examiner as the cited invention during the examination of the A patent document, the C patent document is also quoted with the A patent document.

The first group includes a patent document of the first group and a patent document of the second group. However, the first group searches the document and then classifies the document using the indirect citation relation A document group formed by performing the above process. The second group refers to the patent documents designated by the user or other patent documents stored in the database 130. The second group refers to a group of patent documents for which classification of documents by the document classifying module 150 of the embodiment is not performed .

Accordingly, when the user requests the document classification to be performed on the patent documents of the search result, after the document classification is performed by the document classification module 150, at least one or more groups Can be generated. When the user desires to perform document classification or clustering on other patent documents (hereinafter referred to as "second group document") after the document classification, the documents belonging to the second group of the non- Can be classified and grouped into the classes belonging to the first group using the characteristics of one group (representative document or representative vector).

In order to facilitate understanding, it has been described that the documents belonging to the first group are defined as those in which the document classification using the indirect citation relation is performed, and the documents belonging to the second group are not yet classified or clustered. However, The document belonging to the first group can be classified and clustered according to the classification criteria of the first group even if classification or clustering has already been performed.

In the description of the present invention, the term classification and cluster may be used in combination. However, this may be interpreted in terms of the operations of the document classification module 150 or the document search module 120, It is necessary to note.

Meanwhile, in addition to the reading of the citation relation, the present embodiment can classify the patent documents using the indirect citation relation, which will be described with reference to FIGS. 6 to 8 attached hereto.

FIG. 6 is a diagram showing an example of a document grouping module of the document classifying module according to the present embodiment, FIG. 7 is a diagram for explaining a configuration for deriving an indirect citation relation by the document classifying module according to the present embodiment, FIG. 8 is a diagram for explaining a configuration for grouping similar documents into classified groups by the document classification module according to the present embodiment.

First, referring to FIG. 7, a configuration for deriving an indirect citation relation by the document classifying module 150 of the present embodiment will be described.

The user can obtain information about the indirect citation relation by the document classification module 150 with respect to the document of the searched result or directly designated documents. As shown in FIG. 7, the user can set a period (period A to period B) for the documents to be classified. In this case, among the patent documents to be classified, classification for the documents within the set period is performed do.

In other words, even if a direct citation relationship between patent documents belonging to a set period (a citation relation formed by recording bibliographic information in the document or a citation relation formed by an examiner etc.) is not established, These patent documents may be classified in the same category as indirect citation relationships if there is a relationship between them.

For example, the period set by the user for document analysis and classification is period A to period B, and the patent documents (Base Patent, Patent 5, Patent 6, Patent 7, Patent 8, Patent 9) If the first patent document (Patent 1) outside the set period is cited in the fifth patent document (Patent 5) and the Base Patent, the fifth patent document (Patent 5) and the base patent document There is an indirect citation relation between them.

As another example, if the third patent document (Patent 3) refers directly to the seventh patent document (Patent 7) and the base patent document within the period, the third patent document (Patent 3) The patent document (Patent 7) is categorized into the same category according to the present embodiment since the indirect citation relation is established mutually.

In this case, in the case of FIG. 7, since the indirect citation relation is formed in the base patent document 5 patent document to the ninth patent document (Patents 5 to 9), the representative document or the base patent document .

Then, the user can directly write the category name to easily understand the contents of the category units of the patent documents classified by this method. For example, when a patent document of a classified category shares a solution task (or a corresponding task) called "noise reduction", the category name is written as "noise reduction (ex. , The folder name can be managed as 'noise reduction'.

When the user names the classified categories as resolution tasks 1, 2, 3 and resolution means 1, 2, 3, etc., the display of the categories corresponding to the respective resolution tasks and resolution means An image may be displayed. In this case, the image displayed in the graph can be expressed in a different color or size according to the number of patent documents included in each category, and the image of the sum (or average evaluation value) of the evaluation values of the patent documents included in each category And may be represented in different colors or different sizes.

The user selects a specific category (resolution means 1, resolution means 2, solution means 3, corresponding Assignment 1, corresponding Assignment 2, and Assignment 3) in the UI provided as a result of document classification provided to the user or as a result of document collusion , Information on the patent document having the highest evaluation value given by the representative patent document (base patent document) or the document evaluation module described above is provided to the user.

By this process, the user can perform document classification on the retrieved documents. Furthermore, after classifying documents using indirect citation relations, it is possible to classify and clusters patent documents that are not categorized or classified as other indirect citation relations - possibly belonging to the second group.

Here, the document clustering process may be performed using similarity determination between documents by the document characteristic creation module 160, and the document classification module 150 may classify the first group of patent documents, Classify and cluster patent documents. The document clustering means 152 of the document classifying module 150 determines the degree of similarity between the patent document belonging to the first category of the first category (which may be the representative document of the first category) And judges which category of the first group the patent document belonging to the second group should be categorized by judging.

The document gathering unit 152 may include a representative vector calculating unit 1521 that calculates a representative vector necessary for clustering using a representative document in the classified category or a plurality of documents belonging to the category.

The document clustering unit 152 may include a field clustering unit 1522 for clustering similar documents by fields (or by identification items) constituting the patent document.

The representative vector calculating unit 1521 calculates the representative feature vector of the document based on the occurrence frequency of each keyword from the representative document (the patent document selected by using the base patent document or the evaluation value) in the already formed category or documents belonging to the same category The index file created by the module 160 is used. For example, the representative vector calculating unit 1521 can extract representative keywords having a high frequency from among the keywords appearing in each document, and extracts a plurality of keywords from the index file of each document in order of occurrence frequency Can be selected.

By this selection operation on the keyword distribution diagram as shown in FIG. 9, feature vectors of each document as shown in FIG. 10 can be formed.

For example, in Documents 1, the keyword A is 4.5%, the keyword B is 2.4%, the keyword E is the keyword E, and the keyword E is the keyword B. The representative vector calculating unit 1521 calculates the percentage of each keyword 1.9%, and the keyword D is 1.7%, so that the percentage of occurrence frequency of each keyword can be calculated.

With this process, the percentage of the occurrence frequency of each keyword is calculated for the documents in the category or the representative document (hereinafter referred to as the " category document ").

9 and 10, after this process is performed for the category documents, the percentage of each keyword for the entire category document is summed up, and a predetermined number of specific keywords in the descending order of the percentage of the summed keywords Can be selected as representative keywords.

For example, a value obtained by adding the percentage occupied by each keyword in all ten category documents among the keywords shown in FIG. 9 is a keyword B, a keyword A, a keyword E, a keyword D, a keyword O, a keyword C, a keyword K Keyword A, keyword E, and keyword D can be selected as representative keywords for grouping the selected documents. Then, the feature vector for each document is calculated using the selected representative keyword as a representative vector component. That is, the selected representative keywords are arranged in a high probability distribution order and are selected as the representative vector components. A process of creating a feature vector of each document based on the selected keyword B, the keyword A, the keyword E, and the keyword D is performed. This process is performed on the top four keywords in the index file of each document. However, four representative keywords constituting the components of the representative vector are selected and four keywords having a high frequency are compared with each other to generate feature vectors of the respective documents. However, this is only an example, Can be changed at any time.

If the selected representative keyword is included in each document, the vector component may be set to '1', and if it is not included, it may be set to '0'. However, it is also possible to generate a vector component by assigning weights to the respective keywords instead of 1 and 0 as vector components.

As shown in Fig. 9, the feature vector of each document thus created is set to '1' when the representative keyword is included, and to '0' when it is not included, thereby completing the feature vector of each document.

By this process, the feature vector of Document 1 becomes (1,1,1,1), and the feature vector of Document 2 becomes (1,1,0,1). Although the components of each feature vector are created as 1 or 0, each vector component can be assigned different values depending on the occurrence frequency of each representative keyword.

In the case of using a plurality of category documents, a process of selecting a representative vector (or a center vector) using feature vectors of the respective documents is performed. Here, a representative vector .

In this case, among the respective feature vectors shown in FIG. 9, the feature vector (1,1,1,1) of Document 1 can be a representative vector, and by using the selected representative vector, Documents can be clustered.

By using the representative vector derived from the category document, it is possible to check whether or not a patent document having a predetermined degree of similarity with a specific category is included in the second group. Such similarity can be obtained by using the above- It can also be judged by carrying out on the patent documents of the group.

That is, the similarity between the category document belonging to the predetermined category of the first group and the non-classified document belonging to the second group can be calculated by using the inner product by the feature vector or the representative vector. For example, If the value of the inner product operation through the inner product with the feature vector of the two groups of patent documents falls within a predetermined range, it can be grouped together with the representative vector. That is, the representative vectors can be classified and grouped into categories to which the representative vectors belong.

When the representative vector is A and the feature vector of the document to be compared with the similarity degree is B, the document grouping means 152 sets the inner product value of the vector A and the vector B as | A | ² is divided from '1', the degree of similarity between the document corresponding to the vector A and the document corresponding to the vector B is determined.

However, when the value of the inner product operation between the representative vector and the feature vector of the second group document deviates from the reference value, it is not grouped together with the representative vector, and is used as a document for another group.

According to such similarity degree calculation and determination between the representative vector representing the category and the feature vector of the second group document, the twentieth document P20 belonging to the second group can be grouped into the first group A classification , And the twenty-first document P21 of the second group can be grouped into the B group of the first group.

However, if document classification is performed by the document classification module 150 in addition to the above-described embodiments, the document classification module 150 can select a technical classification code (IPC or F-term) representative of the category have. In this case, the classification and grouping of the second group documents by the document grouping unit 152 uses the technical classification code in addition to the similarity determination described above.

For example, the document grouping unit 152 may use F-terms that record a high frequency for each category that is a result of classifying a document by using the indirect citation relation, term can be determined.

In the case of the F-term, since it is classified according to the solution or solution, if it is used together with the similarity determination using the vectorization of the document, more effective document clustering can be performed.

11 to 15 by the document classifying module 150 and the UI outputting means 112, the document classification module 150 and the UI outputting means 112 are used for classifying the patent documents according to the present embodiment, May be provided to the user.

11 is a first embodiment of a UI for information that can be obtained from the classification and clustering of documents.

After the classification of the patent document is performed by the document analysis system of this embodiment and the other patent documents are clustered by using the classified result, the patent document such as the one shown in FIG. 11 is generated according to the period setting of the user or the setting of the applicant An analysis UI can be provided to the user.

For example, when a user sets his or her setting to "LGE (including representative name)" and competitor is set to "A company", the number of applications by country and the evaluation value Etc. can be shown in the diagram. In particular, the evaluation value given by the document evaluation module 140 may be included, and the evaluation value of the items included in the item may be shown or an average evaluation value of the items included in the item may be displayed.

In addition to this information, a patent index (CPP), a technology influence index (CII), a technology index (TS), an impact measurement index (TII), a technology progress measurement index (TCT) and a technology independence index Can be seen.

Here, the citation index per patent refers to the average number of patent citations of the patent, and may be a value obtained by dividing the number of citations of the patent document by the total number of patents as an item for evaluating the technological progress degree of the enterprise. The technical impact index, for example, is used for evaluating information on the influence of a company's technology on the past five years, showing that the company's patent information is cited in the past five years. CII = The total number of patents / the total number of patents in the previous year).

And, the technology index can be calculated as (CII × number of patents) as an item for quantitatively evaluating the technological influence of the enterprise. The Influence Measurement Index is a measure of the percentage of patents cited in the top 10% or more in a particular technology area in the total number of citations in that technology area. To assess the impact on a specific technology area by company, The number of patents cited above / the total number of patents cited).

The index of technological progress measurement is an evaluation item on the speed of technological progress of the company. It represents the average of the year value corresponding to the median value of the patent year with the quotation patent. ). ≪ / RTI > The technology self-reliance index is an item for evaluating the independence of the technology, and can be calculated in order to obtain the degree to which the patent is cited (the number of patent citations / the total number of citations).

These various kinds of indexes can be calculated by the document classifying module 150 after the classification and clustering of the documents is performed and the results of such operations are displayed by the UI output means 112 as shown in Figs. 11 to 15 A graph or a graph.

Figure 12 is a second embodiment of a UI for information that can be obtained from the classification and clustering of documents. In the case of the second embodiment, the number of patent documents per applicant in the set period is shown in a diagram, and the applicant may be selected by the user.

Then, the average evaluation value of the patent documents corresponding to each period can be shown as W / F, and the user can confirm the position that can become the inflection point of the technology development from the W / F item displayed together with the UI. When the user selects a point having a high W / F as the average evaluation value, the document classifying module 150 and the UI outputting means 112 of the embodiment provide information of the patent documents at that point in a separate UI, A document or a representative document having the highest evaluation value at a time point can be provided in a separate UI.

Figure 13 is a third embodiment of a UI for information that can be obtained from the classification and clustering of documents. 13, a UI including information such as a time set by the user and a patent per person patent (CPP), a technical influence index (CII), a patent per citation index (CPP), and a technical influence index (CII) , And a graph showing the index of patent per person by applicant in the UI in the UI may be further included.

That is, in the UI shown below in FIG. 13, applicants such as Samsung Electronics and Sharp are exemplified as having a high quotation index.

In addition, information on patent activity evaluation, patent activity index, patent portfolio analysis (HHI), and patent diversification index by technology field can be further provided. Patent activity evaluation by technology field is to quantitatively compare patent activity in each field within a selected period , And by filing (or disclosing) applications by technology sector.

The patent activity index indicates the percentage of a specific technology field, which can be calculated as {(number of patents in a specific field / total number of patents of a company) / (total number of patents / total number of patents in all fields)}.

In addition, the patent portfolio analysis index is an item to identify the way in which companies compete in the market, and it can be calculated by the top IPC field of each company, and the technology field competing with the technology field which is dominated by each company can be calculated. For example, the number of applications per inventor indicates the relative evaluation (number of total applications / number of inventors) of the number of applications per inventor, and the number of claims per inventor indicates the relative evaluation of the number of claims acquired per inventor / Company inventors) index, and the effective patent average remaining period may indicate an index of the average remaining period of the patent (the total number of remaining patents of the effective patent / total number of effective patents).

The joint application ratio is an item that evaluates the active degree of the joint research activities and can be calculated as (the number of joint applications / total number of patents).

Figures 14 and 15 are fourth and fifth embodiments of a UI for information that can be obtained from the classification and clustering of documents.

14 and 15 show a graph of a number of quotations per company in a specific period and a UI having a chart of a patent document having a high number of citations. In the display of a patent document having a high number of citations, And the evaluation values given by the evaluation module 140 can be displayed together.

In addition, when a user reports a chart arranged in the order of a high number of citations and selects a number of a specific patent document (application number, registration number, etc.), additional information about the patent document or a corresponding statement can be provided to the user have.

The document classification result or document clustering result provided by the document analysis system of the present embodiment as described above can be stored and shared with other users according to the system setting. In this case, Or it would be very useful on the team.

Claims (10)

A database storing a plurality of patent documents;
A document classifying unit grouping the plurality of patent documents into a plurality of patent documents belonging to the first group so as to belong to the first group; And
And an output unit for outputting information on a plurality of patent documents belonging to a second company in comparison with information on a plurality of patent documents belonging to a first company among a plurality of patent documents belonging to the first group,
The document classifying unit classifies the plurality of patent documents into a first group by using an indirect citation relation appearing through a patent document out of the set period among a plurality of patent documents belonging to a set period
Document analysis system. The method according to claim 1,
The first company is a third party, and the second company is a third party
Document analysis system. The method according to claim 1,
The output unit may output the number of applications per country as information on the plurality of patent documents
Document analysis system. The method according to claim 1,
Wherein the information on the plurality of patent documents includes an evaluation value indicating a frequency rate of the keywords included in the patent document,
Wherein the output unit outputs a sum of the evaluation values or an average of the evaluation values as information on the plurality of patent documents
Document analysis system. The method according to claim 1,
The output unit may output at least one of a patent citation index, a technology influence index, a technology index, an influence measurement index, a technology progress measurement index, and a technology independence index as information on the plurality of patent documents
Document analysis system. The method according to claim 1,
Wherein the output unit outputs information on a plurality of patent documents belonging to the second company in comparison with information on a plurality of patent documents belonging to the first company in accordance with a plurality of technical categories
Document analysis system. delete The method according to claim 1,
Wherein the document classification unit classifies a plurality of patent documents belonging to the second group based on a degree of similarity between a plurality of patent documents belonging to the second group and a plurality of patent documents belonging to the first group
Document analysis system. 9. The method of claim 8,
Wherein a plurality of patent documents belonging to the first group are classified into a plurality of categories,
Wherein the document classification unit classifies a plurality of patent documents belonging to the second group into the plurality of categories based on a degree of similarity between a plurality of patent documents belonging to the second group and a plurality of patent documents belonging to the first group
Document analysis system. 10. The method of claim 9,
Wherein the plurality of categories include a plurality of representative patent documents representing the plurality of categories,
Wherein the document classification unit classifies a plurality of patent documents belonging to the second group and a plurality of patent documents belonging to the second group on the basis of the similarity degree of the plurality of representative patent documents into the plurality of categories
Document analysis system.

Images