KR101199527B1 - System and Method of technology structure analysis based on patent information - Google Patents

Abstract

The present invention relates to a system and method for analyzing technical structure based on patent information, comprising: a clustering module for classifying a patent document according to a right type and clustering the classification results according to similarity of IPC codes, and each cluster by the clustering module. Text mining of the technical classification code extracted from the core technology classification code extraction module and the core technology classification code extraction module for dividing the layers by analyzing the IPC codes of the constituent patent documents, and extracting the technology classification code for each layer. And a technical field derivation module for deriving a technical field by deriving the cowords and matching the derived coords with the patent document information of the upper layer.
Therefore, according to the present invention, it is possible to use the IPC Code (International Patent Classification Code) classification system included in the patent information to create a technical structure, and use the patent document information to quantitatively analyze the field name of the technical structure. Can let

Description

System and Method of technology structure analysis based on patent information}

The present invention relates to a system and method for analyzing technical structures based on patent information, and more particularly, to cluster patent documents according to similarity between rights types and IPC codes. The IPC codes of the patent documents are analyzed to divide the layers, extract the technology classification codes for each layer, and then, by text mining the technology classification codes for each layer, derive the coords, and obtain the coords for each layer above the higher layers. The present invention relates to a technical structure analysis system and method based on patent information which derives a technical field by matching with patent document information.

The knowledge used in the development of new technologies varies greatly in terms of types and forms, but in recent years, attention has been focused on patents that reflect the latest trends in technology.

Patent information has the characteristics of technical information, rights information and management information, and its importance is increasing in the global international competition. As the technical information, it is possible to know the trend of technology development and technical ideas applied to individual patents, and as the right information, it is possible to grasp the scope of rights of individual patents and the degree of nationalization of domestic and foreign rights.

In particular, the patent information as the technical information is used as a very important data for identifying trends and technical characteristics of technological change not only for researchers who perform academic research on technological innovation, but also for engineers who develop actual technologies. Therefore, the development of a new technology discovery system using patent information can be said to be meaningful in terms of using the accumulated knowledge to create new knowledge because the characteristics of the past technologies can be analyzed based on a very large amount of data.

For this reason, a system for analyzing patent trends is needed. Patent trend analysis system uses a patent trend analysis method that does not focus on technical classification by the user. That is, a method of performing patent trend analysis by automatically classifying a technology by IPC classification or the like without using a technical classification of patent data by a user is used. When a user's selection is involved in the technology classification, several main words are selected in advance. Use

The technology classification method of the conventional patent trend analysis system has a problem in relation to patent search and accuracy of the technology classification itself, and in particular, there is a disadvantage in that it cannot accommodate the diversity of technology classification according to the analysis purpose.

In addition, the patent trend analysis system supporting technical classification simply extracts keywords from the entire patent documents by text mining techniques and uses technical classification according to the frequency of the keywords. There is a problem falling.

The method that allows the user to select the main word after the technical classification is to classify the patent literature according to the frequency of the main word, and the accuracy of the technical classification is low and the conceptual word is derived from the technical common sense of the analyst. It cannot be ignored that a difference with a term used in a patent document occurs.

In addition, most of the methodologies for qualitatively constructing a technical structure by finding related patents based on a specific keyword, forming a cluster by finding keywords from an unspecified number of patents, and the technical structure of a cluster composed of a large number of patents There is no methodology for packaging quantitative analysis.

In addition, there have been many studies on how to package related patents in order to construct a specific technology (product), but most of them have the disadvantage of remaining at the level of researching qualitative methodologies depending on expert analysis.

The present invention has been made to solve the above problems, the object of the present invention is to make a technical structure using the IPC Code (International Patent Classification Code) classification system included in the patent information, and to use the patent document information The present invention provides a technical structure analysis system and method based on patent information for deriving a field name of a structure.

Another object of the present invention is to provide a technology structure analysis system and method based on patent information that can overcome the limitations of individual patents in the commercialization of technology by packaging a plurality of patents in technology and product units.

In order to achieve the above objects, according to the present invention, the patent document is classified according to the right type, the clustering module for clustering the classification results according to the similarity of the IPC code, of the patent documents constituting each cluster by the clustering module Core technology classification code extraction module for dividing layers by analyzing IPC codes, extracting technology classification codes for each layer, and text mining the technology classification codes for each layer extracted from the core technology classification code extraction module. And a technical structure analysis system based on patent information including a technical field derivation module for deriving a technical field by matching the derived corridors of each layer with patent document information of an upper layer.

The technical structure analysis system based on the patent information further includes a collection module for collecting patent documents from a source of patent document information, wherein the collection module stores the collected patent documents in a database.

The database includes a patent document information database storing the collected patent documents, a rights type database storing right type information, and an IPC code database storing information on an IPC code.

The technical structure analysis system based on the patent information may further include a technical structure derivation module for deriving a technical structure by matching the technical field derived from the technical field derivation module with the IPC classification system.

The clustering module includes a right type classification unit for classifying the collected patent documents according to a right type, and a data converting unit for vectorizing the patent documents classified with the right type according to an IPC code and clustering according to the similarity of the IPC code. .

The core technology classification code extraction module is a class selector for classifying IPCs having a certain proportion or more of IPCs of a patent document constituting each cluster as class IPCs, and IPCs having a mode value for each class IPC as subclass IPCs. A sub class selector for selecting, a main group selector for selecting an IPC having a mode value for each subclass IPC as a main group IPC, and extracting a plurality of codes by analyzing the main group IPC And a subgroup derivation unit for deriving an IPC code having a mode as a core technology classification code by applying the extracted weight for each codeword.

The class selector selects an IPC having a class IPC ratio of 10% or more as a class IPC.

The technical field deriving module extracts a predetermined number of keywords with high frequency by text mining a description matching the subgroup IPC, and compares the extracted keyword with a catchword index. After deriving, matching the derived code with the patent document information corresponding to the main group IPC, the main group technical field deriving unit for deriving the technical field of the main group of the code having the most frequent value, matching with the main group IPC Text mining the description to extract a certain number of keywords with a high frequency, and compare the extracted keyword with a catchword index (CATCHWORD INDEX) to derive a code, and then the derived code to the sub Corresponding to the patent literature information corresponding to the class IPC, the code having the mode is shown in the technical field of the subclass. The sub-class technical field deriving unit performs text mining on a description matching the sub-class IPC, extracts a predetermined number of keywords, and compares the extracted keywords with a catchword index. After deriving a word, a class description field derivation unit for deriving a technical field of a class having a codeword having a mode value by matching the derived code with the patent document information corresponding to the class IPC, a description matching the class IPC A text mining method extracts a predetermined number of keywords with high frequency, compares the extracted keyword with a catchword index, and derives a codeword, and then uses the derived code to construct the cluster. Clustering apparatus that derives the technical field of the cluster with the mode having the most frequent matched with the information Field includes deriving unit.

In addition, according to the present invention, (a) classifying the patent documents according to the right type, clustering the classification results according to the similarity of the IPC code, (b) IPC of the patent documents constituting each clustered cluster Dividing the layers by analyzing the codes, extracting the technical classification codes for each layer, (c) deriving a codeword by text mining the extracted technical classification codes for each layer, and (d) each derived layer There is provided a technology structure analysis method based on patent information including the step of deriving a technical field by matching the star code with the patent document information of the upper layer.

In the step (c), the method may include extracting a predetermined number of keywords having high frequency by text mining the description classification code of each layer, and deriving a coword by comparing the extracted keyword with a catchword index.

In the step (d), the code for each layer is matched with the patent document information of the upper layer to derive the code having the mode to the technical field of each layer.

As described above, according to the present invention, a technical structure is made by using the IPC Code (International Patent Classification Code) classification system included in the patent information, and quantitative analysis that derives the field name of the technical structure by using the patent document information. Can be enabled.

In addition, since a number of patents are packaged in technology and product units, it is possible to overcome the limitations of use of individual patents in the commercialization of technology.

1 is a view showing a technical structure analysis system based on patent information according to the present invention.
Figure 2 is a block diagram schematically showing the configuration of the technical structure analysis server according to the present invention.
3 shows the clustering module shown in FIG. 2;
4 illustrates the core technology classification code extraction module shown in FIG. 2;
FIG. 5 is a diagram illustrating a technical field derivation module illustrated in FIG. 2.
6 is a flowchart illustrating a technology structure analysis method based on patent information according to the present invention.
7 is an exemplary diagram for explaining a method for extracting a core technology classification code according to the present invention;
8 is an exemplary view for explaining a method for deriving a technical field of each layer according to the present invention.
9 is an exemplary view for explaining a method for deriving a technical structure according to the present invention.

Details of the above-described objects and technical configurations of the present invention and the effects thereof according to the present invention will be more clearly understood by the following detailed description based on the accompanying drawings.

1 is a view showing a technical structure analysis system based on patent information according to the present invention.

Referring to FIG. 1, the technology structure analysis system based on patent information derives clustering and technical fields by collecting patent documents from a patent document information database server 100 and a patent document information database server 100 in which patent document information is stored. The technical structure analysis server 200 is included.

The technical structure analysis server 200 collects and stores a patent document from the patent document information database server 100 and classifies the stored patent document according to a right type and an IPC (International Patent Classification) code type. It creates clusters, derives key technologies using IPC codes of each cluster, and derives technical fields through patent literature information.

The technical structure analysis server 200 may collect patent documents in various ways, but largely by using a patent information system provided by patent offices of each country, by using a patent information system developed by a private company, or by an individual company. It can be divided into using a patent information system built for the purpose of. Examples include patent information systems provided by the Korean Intellectual Property Office (www.kipo.go.kr) and the Korean Patent Information Service (www.kipris.or.kr), www.delphion.com in the United States, and www.patolis.co. In Japan. jp, www.wips.co.kr, www.wisdomain.com of Korea.

2 is a block diagram schematically showing a configuration of a technical structure analysis server according to the present invention, FIG. 3 is a diagram showing a clustering module shown in FIG. 2, and FIG. 4 is a diagram showing a core technology classification code extraction module shown in FIG. FIG. 5 is a diagram illustrating a technical field derivation module illustrated in FIG. 2.

Referring to FIG. 2, the technical structure analysis server 200 may include a collection module 210, a clustering module 230, a core technology classification code extraction module 250, a technology field derivation module 270, and a technology structure analysis module 290. ).

The collection module 210 collects patent documents from a source of patent document information and stores them in a database (not shown). Here, the database includes a patent document information database (not shown), a rights type database (not shown) including an Inventor search term (subsidiary body) of a patent holder, and an IPC code database (not shown) including technical characteristics. do. The right type database stores three types of rights in which the appearance (university), the university and the private sector each hold patent rights, and seven types of rights in which the appearance (university), the university and the enterprise jointly hold patent rights. have. The IPC code database stores information on the IPC code.

The clustering module 230 classifies the patent documents collected by the collection module 210 according to the right type, and clusters the classification results according to the similarity of the IPC codes to generate a cluster.

That is, the clustering module 230 clusters patents having similar code classification schemes based on the IPC Code in consideration of the distribution of rights types because the scope of use varies according to the right types of patents.

Referring to FIG. 3 for the clustering module 230, the clustering module 230 includes a right type classification unit 232 for classifying patent documents according to a right type, and a patent document classified with the right type in an IPC code. The data conversion unit 234 is vectorized and clustered according to the similarity of the IPC code.

The right type classification unit 232 classifies the patent documents collected by the collection module 210 according to the right type classified in the right type DB. That is, the right type classification unit 232 calls the document of the same organization in the collected patent document and appears the technology rights holding type (Year), three rights each of which the university and the private person hold patent rights alone. Types and appearances are classified into seven types of rights that are jointly owned by a university or a company.

The data conversion unit 234 vectorizes the patent documents classified by the right type classification unit 232 according to an IPC code, and generates a cluster by clustering patents having similar IPC codes.

The core technology classification code extraction module 250 extracts the core technology classification code by analyzing the IPC hierarchical structure of each cluster generated by the clustering module 230.

That is, the core technology classification code extraction module 250 analyzes the IPC codes of the patent documents constituting each cluster, selects IPCs according to the frequency, divides the layers, and analyzes the IPC hierarchical structure. Derive the code.

In other words, the core technology classification code extraction module 250 analyzes IPC codes of patent documents constituting each cluster and divides the hierarchy into IPCs of classes, subclasses, main groups, and subgroups. At this time, the core technology classification code extraction module 250 derives the IPC code of the subgroup as a core technology classification code.

Referring to FIG. 4 for the core technology classification code extraction module 250 performing the above role, the core technology classification code extraction module 250 includes a class selector 252, a subclass selector 254, and a main module. The group selector 256 and the sub group selector 258 are included.

The class selector 252 selects, as a class IPC, an IPC having an IPC specific ratio of a certain ratio or more (for example, 10% or more) constituting each cluster.

The subclass selector 254 selects an IPC having a mode value for each class IPC selected by the class selector 252 as a subclass IPC.

The main group selector 256 selects an IPC having a mode value for each subclass IPC selected by the sub class selector 254 as a main group IPC.

The subgroup deriving unit 258 extracts a plurality of cowords by analyzing the main group IPC selected by the main group selecting unit 256, and applies a weight value for each of the selected cowords to form a mode value. The IPC code with is derived as the core technical classification code.

That is, the subgroup derivation unit 258 analyzes the description of the main group IPC, extracts cowords representing the main group IPC, and includes the most extracted subgroup IPC code. Is selected and adopted as the Seed-Tech classification code. Here, the Description means International Patent Classification User's Guide Description.

The technical field derivation module 270 performs text mining on the technology classification code of each layer selected by the core technology classification code extraction module 250 to derive a co-word, and the derived code of each layer. The technical field is derived by matching the word with the patent document information of the upper layer.

That is, the technical field derivation module 270 performs text-mining on the IPC Main Description corresponding to the IPC of the lower layer constituting each cluster, extracting the top 30 keywords with high frequency, and extracts the extracted keywords from the CATCHWORD INDEX. In contrast, Co-words are derived. Then, the technical field derivation module 270 matches the patent document information constituting the IPC code of the upper layer with respect to the lower layer and the derived code so that the code having the mode is the highest level. To derive. Here, the CATCHWORD INDEX refers to a collection of international patent classifications configured to display the technical terms in alphabetical order so as to know what patent classification corresponds to when searching for related terms.

Through the above process, the technical field derivation module 270 derives the technical field for the main group, subclass, class, and cluster hierarchy.

Referring to FIG. 4 for the technical field derivation module 270 performing the above role, the technical field derivation module 270 may include a main group technical field derivation unit 272 and a subclass technical field derivation unit 274. The class description field derivation unit 276 and the cluster description field derivation unit 278 are included.

The main group technical field derivation unit 272 extracts a predetermined number of keywords with high frequency by text mining a description matching the subgroup IPC, and extracts the extracted keywords from a catchword index (CATCHWORD INDEX). Contrast is derived. Then, the main group technical field deriving unit 272 derives a word having a mode value to the technical field of the main group by matching the derived code with the patent document information corresponding to the main group IPC.

The sub-class technical field deriving unit 274 extracts a predetermined number of keywords having high frequency by text mining a description matching the main group IPC, and extracts the extracted keywords from a catchword index and a catchword index. Contrast is derived. Then, the sub-class technical field deriving unit 274 matches the derived code with patent document information corresponding to the sub-class IPC to derive a code having a mode value into the technical field of the sub-class.

The class description field deriving unit 276 performs text mining on a description matching the subclass IPC, extracts a predetermined number of keywords, and compares the extracted keyword with a catchword index. To derive the coword. Then, the class technical field derivation unit 276 matches the derived code with the patent document information corresponding to the class IPC to derive the code having the mode to the technical field of the class.

The cluster technical field derivation unit 378 extracts a predetermined number of keywords with high frequency by text mining a description matching the class IPC, and compares the extracted keyword with a catchword index. Derive a coinword Then, the cluster technical field derivation unit 278 matches the derived code with the patent document information constituting the cluster to derive the code having the mode into the technical field of the cluster.

The technical structure analysis module 290 derives the technical structure by matching the technical field derived from the technical field derivation module 270 with the IPC classification system.

6 is a flowchart illustrating a method for analyzing technology structure based on patent information according to the present invention, FIG. 7 is an exemplary view for explaining a method for extracting a core technology classification code according to the present invention, and FIG. 8 is a description of each layer according to the present invention. 9 is an exemplary diagram for describing a method for deriving a field, and FIG. 9 is an exemplary diagram for describing a method for deriving a technical structure according to the present invention.

Referring to FIG. 6, the technical structure analysis server collects patent documents (S600), classifies the collected patent documents according to a right type, and clusters the classification results according to similarity of IPC codes (S602).

That is, the technical structure analysis server classifies the technical right types as shown in Table 1.

group Subgroup Contents A A-1
A-2
A-3 When Appearance Applied for Patent
If a university has applied for a patent
When a private patent is filed B B-1
B-2
B-3
B-4
B-5
B-6
B-7 When appearance appears jointly for patent
If a university jointly filed a patent
When a private patent is filed jointly
When a patent is filed by appearance and private cooperation
When patents are jointly filed
When a patent is filed jointly by a university and a private company
Applied patents and joint patent applications between universities and the private sector

The technical structure analysis server then clusters the results classified according to the right type according to the similarity of the IPC codes.

After performing the step S602, the technical structure analysis server analyzes the IPC code for each clustered cluster to divide the layer, and analyzes the IPC hierarchical structure to extract the core technical classification code (S604).

A method of extracting the core technology classification code by the technology structure analysis server will be described with reference to FIG. 7.

Referring to FIG. 7, C07 (146), C08 (73), B01 (38), and C09 (35), in which the technical structure analysis server has an IPC weight ratio of 10% or more of the patent documents constituting each cluster. Is selected as the class IPC.

Then, the technical structure analysis server selects C09K (30), A61K (26), B01J (24), etc., which have the mode for each of the selected class IPCs, as subclass IPCs.

The technical structure analysis server selects C09K011 (27), C07D403 (18), A23L009 (12), and the like having a mode value according to the selected subclass IPCs as a main group IPC.

Then, the technical structure analysis server checks the contents of the selected main group IPC and extracts a plurality of words. Here, the codeword may include a keyword or a key clause.

The technical structure analysis server derives the IPC code having the mode, that is, C09K011-80 as the core technical classification code by applying the extracted weight for each codeword.

After performing the step S604, the technical structure analysis server text mines the extracted technical classification code of each layer to derive a code (S606), and the derived code for each layer and the patent document information of the upper group The technical field is derived by matching (S608).

A method of deriving the technical field by the technical structure analysis server will be described with reference to FIG. 8.

Referring to FIG. 8, when the core technology classification code is C09K011-80, the technology structure analysis server performs text mining on IPC Main Description (Main Group Level) corresponding to C09K011-80, and generates the top 30 keywords with high frequency. Extracts a coword by comparing the extracted keyword with a catchword index. Here, the extracted code may be a fluorescent substance, emission luminance, and the like.

 The technical structure analysis server then derives a technical field of 'fluorescent material' by matching the extracted code with the information of C09K011 patent documents of the upper layer. At this time, the technical field may be preset by an expert according to the frequency of the code in the patent document information, the degree of matching between the code and the patent document information.

Then, the technical analysis server performs text mining on IPC Main Description (Main Group Level) corresponding to C09K011, extracts a coword by comparing the top 30 keywords with a high frequency with a catchword index, and extracts the coward. And 'organic phosphors' are derived by matching the C09K patent document information included in the upper group of the hierarchy. The technical analysis server derives a technical field of 'Flat Panel light emitting material' and display light emitting material using the above method.

After performing S608, the technical structure analysis server derives the technical structure by matching the derived technical field with the IPC classification scheme (S610).

A method of deriving the technical structure by the technical structure analysis server will be described with reference to FIG. 9.

The technical structure analysis server derives a Contours-Tech Domain by matching the subgroup IPC Code and the main group code.

In addition, the technical structure analysis server derives Adjacent-Tech by matching the main group IPC Code and the sub-group's code.

In addition, the technical structure analysis server derives Class-Tech by matching subclass IPC Code and the code of the main group.

In addition, the technical structure analysis server derives Cluster-Tech by matching the class IPC Code and the subclass code.

In other words, the technical structure analysis server compares and analyzes the co-word of the IPC lower level and the patent document information of the IPC upper level and analyzes the Seed-Tech (IPC Sub Group Level), Adjacent-Tech (IPC Main Group Level), and Contours- We derive an analytic model classified into Tech (IPC Sub Class level), Class-tech (IPC Class Level), and Cluster-Tech.

Thus, those skilled in the art will appreciate that the present invention may be embodied in other specific forms without departing from the spirit or essential characteristics thereof. It is therefore to be understood that the embodiments described above are to be considered in all respects only as illustrative and not restrictive. The scope of the present invention is defined by the appended claims rather than the detailed description and all changes or modifications derived from the meaning and scope of the claims and their equivalents are to be construed as being included within the scope of the present invention do.

100: patent document information database server 200: technical structure analysis server
210: data collection module 220: clustering module
232: right type classification unit 234: data conversion unit
250: core technology classification code extraction module 252: class selector
254: sub class selector 256: main group selector
258: sub-group selector 270: technical field derivation module
272: Derivation part of the main group technical field 274: Derivation part of the subclass technical field
276: Class deduction field 278: Cluster technical field derivation unit
290: Technical Structure Analysis Module

Claims (13)

A clustering module classifying the patent documents according to the right type and clustering the classification results according to the similarity of the International Patent Classification (IPC) code;
After classifying IPCs having a certain proportion or more of IPCs of the patent documents constituting each cluster by the clustering module as class IPCs, and classifying IPCs having mode values for each class IPC as subclass IPCs, The IPC having a mode is selected as a main group IPC, the main group IPC is analyzed, a plurality of codes are extracted, and the extracted weight is applied to the IPC code having a mode as a core technology classification code. A core technology classification code extraction module for extracting a technology classification code for each layer to be derived; And
The technology classification code of each layer extracted by the core technology classification code extraction module is text mined to derive cowords, and the derived technical field is matched with the patent document information of the upper layer. Deriving technical field derivation module;
Technical structure analysis system based on the patent information comprising a.
The method of claim 1,
Further comprising a collection module for collecting patent documents from a source of patent document information,
The collection module is a technical structure analysis system based on patent information, characterized in that for storing the collected patent documents in a database.
The method of claim 2,
The database includes a patent document information database storing the collected patent documents, a rights type database storing rights type information, and an IPC code database storing information on an IPC code.
The method of claim 1,
The technical structure analysis system based on patent information further comprising a technical structure derivation module for deriving a technical structure by matching the technical field derived from the technical field derivation module and the IPC classification system.
The method of claim 1,
The clustering module,
A right type classification unit for classifying the patent document according to a right type;
And a data conversion unit for vectorizing the patent documents classified by the right type according to the IPC code, and clustering the patent documents according to the similarity of the IPC code.
The method of claim 1,
The core technology classification code extraction module,
A class selection unit for selecting IPCs having a specific ratio of IPCs of the patent documents constituting each cluster into class IPCs;
A sub class selector for selecting an IPC having a mode for each class IPC as a sub class IPC;
A main group selector configured to select an IPC having a mode value for each subclass IPC as a main group IPC; and
And a subgroup deriving unit for extracting a plurality of codes by analyzing the main group IPC, and deriving an IPC code having a mode as a core technology classification code by applying the extracted weight for each codeword. Technical structure analysis system based on patent information.
The method according to claim 6,
The class selection unit is a technology structure analysis system based on patent information, characterized in that the IPC portion of the patent document is more than 10% of the IPC screening class IPC.
The method of claim 1,
The technical field derivation module,
The text mining description matching the subgroup IPC extracts a predetermined number of keywords with a high frequency, compares the extracted keyword with a catchword index, and then derives a coword, and then extracts the code. A main group technical field derivation unit for matching a word with patent document information corresponding to the main group IPC to derive a code field having a mode value of the main group;
After text mining the description matching the main group IPC, a predetermined number of keywords having a high frequency are extracted, and the extracted keywords are compared with a catchword index to derive a coword, and then subtract the derived coword. A subclass technical field derivation unit for deriving a technical field of a subclass with a code having a mode value matching the patent document information corresponding to the class IPC;
The text mining description matching the subclass IPC extracts a predetermined number of keywords having a high frequency, and compares the extracted keyword with a catchword index to derive a codeword and then classify the derived codeword. A class technical field derivation unit for deriving a technical field of a class having a codeword having a mode value matching the patent document information corresponding to the IPC;
The text mining description matching the class IPC extracts a predetermined number of keywords with a high frequency, compares the extracted keyword with a catchword index, derives a codeword, and then decodes the derived codeword. And a cluster technology field derivation unit for deriving a technology field of a cluster having a codeword matching the patent document information constituting the patent document information.
In the technical structure analysis server to analyze the technical structure based on the patent information,
(a) classifying the patent documents according to the type of rights and clustering the classification results according to the similarity of the IPC codes;
(b) classifying IPCs whose IPC specific gravity is greater than or equal to a predetermined ratio of the patent documents constituting the clustered clusters as class IPCs, and classifying IPCs having mode values for each class IPC as subclass IPCs; The IPC having a mode for each class IPC is selected as a main group IPC, the main group IPC is analyzed, a plurality of codes are extracted, and the weighted value of each extracted code is applied to core an IPC code having a mode. Extracting a technical classification code for each layer, which is derived as a technical classification code;
(c) text mining the extracted technical classification codes of each layer to derive a code; and
(d) deriving a technical field by matching the derived corridors of each layer with patent document information of an upper layer;
Technical structure analysis method based on the patent information, including.
10. The method of claim 9,
After the step (d), the technical structure analysis method based on patent information further comprising the step of deriving a technical structure by matching the derived technical field and IPC classification scheme.
10. The method of claim 9,
Before the step (a), the method for analyzing the technical structure based on the patent information further comprising the step of collecting the patent document from the source of the patent document information and storing in the database.
10. The method of claim 9,
Step (c) may include: mining a predetermined number of keywords having a high frequency by text mining the technology classification code for each layer; and
And deriving a wordword by comparing the extracted keyword with a catchword index.
10. The method of claim 9,
In the step (d), the method of analyzing the technical structure based on the patent information, wherein a code having a mode value is derived to the technical field of each layer by matching the code of each layer with the patent document information of the upper layer, respectively.

Images