JPWO2008053949A1 - Document group analyzer - Google Patents

Abstract

In the document analysis apparatus, a text data acquisition unit 101 that acquires a plurality of technical documents, a document vector acquisition unit 102 that calculates a weighting amount of each index word for each acquired technical document, and each acquired technical document as a subject, Using the weighting amount of each index word, perform a factor analysis with each index word as an observation variable, calculate the factor loading for each factor for each index word, and for each factor for each technical document The factor calculation unit 103 that calculates the factor score and the factor loading of each index word are used to determine the attribution factor of each index word, and the factor of each technical document is determined using the factor score of each technical document And an attribution factor determination unit 104. As a result, it is possible to accurately analyze what concept or characteristic group or document group exists in a plurality of documents.

Description

  The present invention relates to a technique for analyzing technical literature, and more particularly to a technique for analyzing what concept or feature exists in a plurality of documents.

In order to analyze the contents of a plurality of technical documents, one that classifies technical documents into a plurality of clusters is known. For example, Japanese Patent Laying-Open No. 2005-92443 (Patent Document 1) adds weights to each word obtained by performing morphological analysis on a searched technical document, vectorizes each technical document, and determines the direction of the vector. Combining technical documents that are close to each other in one cluster. Then, important words are extracted for each cluster.
JP 2005-92443 A

However, in such classification based on conventional cluster analysis, it may be difficult to accurately grasp the concept or feature of the technical field represented in a plurality of technical documents to be analyzed. For example, in JP-A-2005-92443 (Patent Document 1), classification is performed on the scale of the closeness of the vector direction. It becomes a cluster, and the characteristics of each cluster are not always clear. Even if an important word is extracted for each cluster, it is difficult to grasp the difference between clusters having similar extracted important words, and the validity of the classification itself is questioned.
In other words, the technique disclosed in Patent Document 1 does not take into consideration that the analyst grasps the characteristics of the classified cluster. Therefore, if an analyst tries to grasp the characteristics of the classified cluster, the analyst is forced to read each technical document belonging to the cluster, and takes a long time for the analysis.

  Therefore, the present invention has been made in view of the above circumstances, and an object of the present invention is to provide a technical field represented by a plurality of technical literatures when an analyst analyzes technical literatures while maintaining objectivity of analysis. This is to make it possible to grasp what concept or feature has the.

(1) In order to solve the above-described problem, one embodiment of the present invention is applied to a document group analysis apparatus that analyzes text data.
The document group analyzing apparatus includes: text data acquisition means for acquiring a plurality of technical documents represented by text data; and weighting amount calculation means for calculating a weighting amount of each index word for each of the acquired technical documents. The obtained technical documents are used as subjects, and the weight analysis of each index word is used to perform a factor analysis using each index word as an observation variable. For each index word, factor loading is performed for each factor. For each of the technical documents, the calculation means for calculating the factor score for each factor and the factor loading of each index word are used to determine the attribution factor of each index word. The attribution factor determining means for determining the attribution factor of each technical document using the factor score of the technical document and the index word or index word group belonging to the same factor, respectively, the technical literature or technical sentence belonging to each corresponding factor Together with the group of data, characterized by comprising output means for outputting for each factor, a.

  According to one aspect of the present invention, an index word and a plurality of technical documents can be attributed to each factor, so that characteristics or concepts of the technical field are grasped from the index word through linguistic information understandable to the analyst. It becomes possible. In addition, from a plurality of technical documents, it is possible to grasp the characteristics or concepts of the technical field from the specific tendency of the bibliographic information included in each technical document. In the present invention, the factor analysis technique is used, but it does not require the subjective or arbitrary judgment of the analyst, such as “interpretation of factors” in ordinary factor analysis. This is because each index word is used as an observation variable. That is, since the observation variable itself represents the content, if the observation variable belonging to the factor is determined based on the factor loading, the content of each factor can be simply indicated using the observation variable.

(2) In the document group analyzing apparatus,
The attribution factor determination means selects, for each index word, a factor having the maximum factor loading using the calculated factor loading, and specifies the selected factor as an attribution factor of the index word. At the same time, for each technical document, the factor score having the maximum factor score may be selected using the calculated factor score, and the selected factor may be specified as the attribution factor of the technical document.

  By belonging to a factor most relevant to each index word or each technical document, it becomes possible to best explain the factor. As a result, it becomes possible to grasp the features and concepts of the technical field more clearly.

(3) In the document group analyzing apparatus,
The appearance frequency of index words included in the plurality of documents is obtained, the importance of each index word is calculated using the appearance frequency, and a predetermined number of index words with the highest importance are extracted using the calculated importance An important index word extracting unit, and the weighting amount calculating unit may obtain a weighting amount of a predetermined number of index words having higher importance as a weighting amount of each index word.

  By extracting and analyzing only the index words that represent the characteristics of the entire technical document group as important index words from multiple technical documents, it is possible to grasp the characteristics or concepts of the technical field more clearly. is there. On the other hand, it is possible to improve the efficiency of the analysis process by narrowing down the index words in advance.

(4) In the document group analyzing apparatus,
For each of the factors, a factor evaluation value calculation unit that calculates a factor evaluation value indicating a technical evaluation of the factor is provided, and the output unit uses the technical document or technical document as data of the technical document or technical document group. The group factor evaluation value may be output.

  By the said structure, the relative comparison between each factor by a factor evaluation value can be performed. As a result, it is possible to grasp the relative positional relationship between the technical elements represented by the technical literature belonging to the factors, and further classify the important technical elements from those that do not. You can also.

(5) In the document group analyzing apparatus,
The technical document is a patent document including a patent publication and a patent publication, and for each of the factors, the technical document or the technology belonging to the factor, and a progress information acquisition unit that acquires the progress information of each patent document Factor evaluation value calculation means for calculating a factor evaluation value indicating technical evaluation of the factor using progress information of the document group, and the output means, as the technical literature or the technical literature group data, The factor evaluation value may be output.

  In one aspect of the present invention, factors can be evaluated based on a patent gazette. Therefore, by using examination progress information included in patent information related to the patent gazette, highly accurate factor evaluation can be performed. It becomes possible.

(6) In the document group analyzing apparatus,
Document number judging means for judging the number of documents of technical documents or technical document groups belonging to each factor, and the factor evaluation value calculating means for each factor is the number of technical documents or technical document groups belonging to the factor. A first index with a predetermined weight is calculated, a second index obtained by indexing the progress information of the technical documents or technical document groups belonging to the factor is calculated, and the calculated first index and second index are used. The factor evaluation value of the factor may be calculated.

  By determining the number of technical documents, it is possible to grasp the share of factors in a population composed of technical documents. In addition, by applying a predetermined weight to the number of documents, it is possible to evaluate factors that take into account economic aspects and technical competitiveness. Then, quantitative and objective evaluation of factors can be performed by indexing the progress information. As a result, it is possible not only to grasp the share and economic value of the technical elements in the technical field, but also to quantitatively grasp the share and economic value of these technical elements after quantification.

(7) In the document group analyzing apparatus,
The progress information includes the number of citations from other companies, the number of oppositions to be patented, the number of requests for patent invalidation trials, the presence or absence of examination requests, and the presence or absence of registration of patent right setting. The number of documents is a value weighted using at least one of the total value of the number of citations from other companies, the total value of the number of oppositions to be patented, and the total number of requests for patent invalidation, The second index, which is an index of historical information, is the total number of citations from other companies, the total number of oppositions to be patented, the total number of requests for patent invalidation trials, the examination request rate, and the registered assessment rate. It may be a value obtained by indexing at least one of the above.

  It is possible to evaluate factors that take into account the influence of patents that can be an obstacle to patent acquisition and technology development of other companies. In addition, it is possible to perform factor evaluation that takes into consideration the applicant's willingness to acquire rights and the examiner's evaluation. Thereby, since fairness and appropriateness of factor evaluation can be ensured, as a result, the relative positional relationship between the technical elements and the result of importance can be grasped fairly and appropriately.

(8) In the document group analyzing apparatus,
The output means may calculate the factor evaluation value for each factor and for each applicant.

  For each factor, the rank and share of applicants of patent publications belonging to the factor can be grasped. As a result, it is possible to grasp the characteristics of a predetermined technical field from the viewpoint of a competitive state in a company that is a development subject.

  (9 and 10) According to another aspect of the present invention, a data analysis method including the same steps as the methods executed by each of the above apparatuses and a computer that executes the same processes as the processes executed by the respective apparatuses are executed. It is a program that can. This program may be recorded on a recording medium such as FD, CDROM, or DVD, or may be transmitted / received via a network.

(11) In the document group analyzing apparatus,
The technical literature is a patent document including a patent publication and a patent publication,
For each acquired patent document, means for acquiring a patent score that individually evaluates the value of the patent document;
Each of the factors may be provided with factor evaluation value calculating means for calculating a factor evaluation value indicating a technical evaluation of the factor using the patent score of a patent document belonging to the factor.

  By using a patent score that individually evaluates the value of each patent document, a factor evaluation value that reflects the value of the patent document belonging to each factor can be calculated. As a result, it becomes possible to grasp the features and concepts of the technical field more clearly.

(12) In the document group analyzing apparatus,
The factor evaluation value calculation means includes
For each factor, a patent score that is equal to or higher than a predetermined threshold is selected from among the patent scores of patent documents belonging to the factor, and a value obtained by tabulating the selected patent scores may be calculated as the factor evaluation value. Good.

  By subtracting values below the threshold with the value above the specified threshold being the target of aggregation, it is possible to prevent a high number of factors with few important patents from having a high number of less important patents, even if there are many cases. . As a result, an appropriate factor evaluation value can be calculated, and it becomes possible to grasp the characteristics or concepts of the technical field more clearly.

(13) In the document group analyzing apparatus,
The patent score is preferably a value standardized in a document group of a population including a factor for which the factor evaluation value is calculated.

  By calculating the standard value in the population and calculating the factor evaluation value, it is possible to improve the accuracy of relative comparison between different factors. As a result, it becomes possible to grasp the features and concepts of the technical field more clearly.

(14) In the document group analyzing apparatus,
The patent score means that the patent documents are classified into groups for each technical field and every predetermined period, and for each classified group, the progress information of the patent documents belonging to the group is used, It may be a calculated value.

  By classifying into groups for each technical field and every predetermined period, and calculating the patent score using the progress information for each classified group, to correct the bias of the progress information due to the difference in the technical field and the application time, An accurate patent score can be calculated. As a result, an appropriate factor evaluation value can be calculated, and it becomes possible to grasp the characteristics or concepts of the technical field more clearly.

The figure which shows the hardware constitutions of the document group analysis apparatus which concerns on one Embodiment of this invention. FIG. 2A is a flowchart for explaining a procedure of factor extraction processing in the document group analysis apparatus, and FIG. 2B is a flowchart for explaining a processing procedure for calculating a factor progress information score as document or document group data. . Explanatory drawing regarding an example of the method of acquiring the text data of the some document used as analysis object. Explanatory drawing regarding an example of the method of acquiring the text data of the some document used as analysis object. Explanatory drawing regarding an example of the method of acquiring the text data of the some document used as analysis object. Factor load amount of each index word calculated in the embodiment of the present invention. Factor scores for each publication calculated in the examples of the present invention. An index word belonging to each factor extracted in the embodiment of the present invention and a gazette belonging to each factor. The patent impact index calculated based on the some parameter | index of each factor calculated in the Example of this invention. A plurality of indices of each factor calculated in the embodiment of the present invention, and a progress information index calculated based on the indices. Factor progress information score (for each factor) calculated in the example of the present invention. The example which illustrated the factor progress information score calculated in the Example of this invention. The factor progress information score for every factor and every applicant calculated in the Example of this invention. In the Example of this invention, the example which further classified the publication group which belongs to each factor for every applicant, and illustrated the factor progress information score for every factor and every applicant. The functional block diagram of the document group analyzer which concerns on the modification of the said embodiment. The flowchart which shows the procedure of the calculation process of the technical element score in the said modification. The figure which showed distribution of the technical element score of the said modification, and the factor progress information score of the said Example in relation to the number of publications. The example which illustrated the technical element score for every factor and every applicant in the said modification. The figure which showed schematically an example of the data structure of the progress information utilized in the said modification. The figure which showed schematically an example of the data structure of the content information utilized in the said modification. The flowchart which showed the procedure of the calculation process of the patent score in the said modification. The flowchart which shows the detail of the process which calculates the evaluation value of each patent data in the said modification.

Explanation of symbols

1: processing device, 2: input device, 3: recording device, 4: output device, 101: text data acquisition means, 102: document vector acquisition means, 103: factor load and factor score calculation means, 104: attribution factor determination means

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
<1. Configuration of document group analyzer>
FIG. 1 is a diagram showing a hardware configuration of a document group analysis apparatus according to an embodiment of the present invention. The document group analysis apparatus according to the present embodiment includes a processing device 1 having a CPU (central processing unit) and a memory (recording device), an input device 2 as input means such as a keyboard (manual input device), and data of a document group. And a recording device 3 which is a recording means for storing work conditions and processing results by the processing device 1, and an output device 4 which is an output means for displaying or printing data of documents or document groups belonging to the extracted factors. It consists of a computer device.

The processing apparatus 1 includes a text data acquisition unit 101, a document vector acquisition unit 102, a factor calculation unit 103, an attribution factor determination unit 104, a document number determination unit 201, a progress information reading unit 202, an index calculation unit 203, and a patent impact calculation unit 204. A progress information calculating unit 205 and a score calculating unit 206.
Here, in this embodiment, each functional unit (text data acquisition unit 101, document vector acquisition unit 102, factor calculation unit 103, attribution factor determination unit 104, document number determination unit 201, progress information reading unit 202 of the processing apparatus 1 is described. The case where the index calculation unit 203, the patent impact calculation unit 204, the progress information calculation unit 205, and the score calculation unit 206) are realized by software is taken as an example.
Specifically, in the memory of the processing device 1, a program (text data acquisition program, document vector acquisition program, factor calculation program, attribution factor determination program, document number determination program, progress, etc. for realizing the function of each functional unit Information reading program, index calculation program, patent impact calculation program, progress information calculation program, and score calculation program) are stored. And the function of each function part of processing device 1 is realized when CPU runs the above-mentioned program memorized by memory.

The recording device 3 includes a condition recording unit 31, a work result storage unit 32, a document storage unit 33, and the like. The document storage unit 33 includes document group data obtained from an external database or an internal database. The external database means, for example, a document database such as IPDL of a patent digital library serviced by the Japan Patent Office or PATOLIS (registered trademark) serviced by Patrice Co., Ltd. Also, the internal database is a database that stores data such as patent JP-ROM that is sold on its own, FD (flexible disc), CD (compact disc) ROM, MO (magneto-optical disc), DVD that stores documents. A device that reads from a medium such as a (digital versatile disk), a device such as an OCR (optical character reader) that reads a document that has been output or handwritten on paper, and a device that converts the read data into electronic data such as text And so on.
In this embodiment, as the analysis target document, mainly published patent gazette, published patent gazette, patent publication gazette, patent invention specification, published patent gazette, re-published patent gazette, patent trial request publication, published patent gazette English abstract, published utility New model bulletin, published utility model specification, published utility model gazette, republished utility model gazette, public utility model gazette, utility model registration gazette, registered utility model gazette, registered utility model specification, utility model trial notice, public technical report However, the present invention is not limited to this, and it is possible to analyze general technical literature such as technical papers, magazines dealing with technology, and books.

  As a communication means for exchanging signals and data among the processing device 1, input device 2, recording device 3, and output device 4, it may be directly connected by a USB (Universal Serial Bus) cable or the like, or LAN (local It may be transmitted / received via a network such as an area network, or may be via a medium such as an FD, CDROM, MO, or DVD that stores a document. Alternatively, a part or a combination of these may be used.

<1-1. Details of Input Device 2>
Next, the configuration and function of the document group analysis apparatus will be described in detail.
In the input device 2, the text data acquisition condition of the analysis target document group, the document vector acquisition condition, the factor loading amount and the factor score calculation condition, the attribution factor determination condition, the factor progress information score calculation condition described later, and the output condition Etc. are accepted. These input conditions are sent to and stored in the condition recording unit 31 of the recording device 3.

<1-2. Details of Processing Device 1>
The text data acquisition unit 101 acquires data of a document group to be analyzed from the document storage unit 33 of the recording device 3 in accordance with the acquisition conditions of the text data input by the input device 2. For example, text data is acquired for I documents belonging to one cluster among the clusters obtained as a result of cluster analysis of document groups extracted under a certain condition based on the similarity. The acquired text data is directly sent to the document vector acquisition unit 102 and used for processing there, or is sent to the work result storage unit 32 of the recording device 3 and stored therein.

The document vector acquisition unit 102 calculates I document vectors based on the text data of I documents acquired by the text data acquisition unit 101 in accordance with the document vector acquisition conditions input by the input device 2. When the number of index words is J, this document vector is a J-dimensional vector having the weighting amount z _ij of the index word j in each document i as a vector element. The calculated document vector is directly sent to the factor calculation unit 103 and used for processing there, or is sent to the work result storage unit 32 of the recording device 3 and stored therein.

The factor calculation unit 103 follows the factor load amount a _jk and the factor based on the vector element z _ij of the document vector calculated by the document vector acquisition unit 102 according to the factor load amount and factor score calculation conditions input by the input device 2. A score f _ik is calculated. Here, k is a factor number, a factor loading a _jk is calculated for each factor for each index word j, and a factor score f _ik is calculated for each factor for each document i. The calculated factor load a _jk and factor score f _ik are directly sent to the attribution factor determination unit 104 and used for processing there, or sent to the work result storage unit 32 of the recording device 3 and stored therein.

The attribution factor determination unit 104 determines the attribution factor of each index word j based on the factor loading a _jk calculated by the factor calculation unit 103 in accordance with the attribution factor determination condition input by the input device 2, and factor score The attribution factor of each document i is determined based on f _ik . The determined attribution factor is directly sent to the document number determination unit 201 and the progress information reading unit 202 and used for processing there, or is sent to the work result storage unit 32 of the recording apparatus 3 and stored therein. Then, according to the output condition input by the input device 2, the index word j belonging to the same factor is output by the output device 4 together with the data of the corresponding document i.

The document number determination unit 201 to the score calculation unit 206 calculate the factor progress information score in accordance with the factor progress information score calculation condition input by the input device 2.
Based on the attribution factor of each document determined by the attribution factor determination unit 104, the document number determination unit 201 sets a document group for each factor or a document group for each factor and each applicant as a score calculation target document group. Read and determine the number of documents.
The progress information reading unit 202 reads the progress information of each document from the document storage unit 33 of the recording device 3 for the score calculation target document group.
The index calculation unit 203 calculates an index based on the progress information read by the progress information reading unit 202 for the score calculation target document group.
The patent impact calculation unit 204 calculates the patent impact of the score calculation target document group based on the number of documents determined by the document number determination unit 201 and the index calculated by the index calculation unit 203.
The progress information calculation unit 205 calculates a progress information index based on the index calculated by the index calculation unit 203 for the score calculation target document group.
The score calculation unit 206 calculates a factor progress information score for the score calculation target document group based on the patent impact calculated by the patent impact calculation unit 204 and the progress information index calculated by the progress information calculation unit 205.
The work result by each functional unit is sent to and stored in the work result storage unit 32 of the recording device 3.

<1-3. Details of Recording Device 3>
In the recording device 3, the condition recording unit 31 records information such as conditions obtained from the input device 2, and sends necessary data based on a request from the processing device 1. The work result storage unit 32 stores the work result of each component in the processing device 1 and sends necessary data based on a request from the processing device 1. The document storage unit 33 stores and provides necessary document group data obtained from an external database or an internal database based on a request from the input device 2 or the processing device 1. When storing patent document data, the document storage unit 33 preferably stores the bibliographic information (such as the name of the applicant) and the progress information (information such as the examination request).

<1-4. Details of Output Device 4>
The output device 4 outputs, for each factor, the document and index word for which the attribution factor is determined by the attribution factor determination unit 104 of the processing device 1. The output device 4 includes a display unit such as a display device, and displays a correspondence table between documents and / or index words and factors, a factor evaluation value of a document or a document group calculated for each factor, and the like. The form of output is not limited to display on the display unit, but may be printing on a print medium such as paper or transmission to a computer apparatus on a network via a communication unit.

<2. Factor extraction process>
FIG. 2A is a flowchart for explaining the procedure of factor extraction processing in the document group analysis apparatus. The document group analysis apparatus of this embodiment extracts factors from a plurality of documents to be analyzed using a factor analysis technique.

<2-1. Acquisition of text data>
In the document group analysis apparatus of this embodiment, the text data acquisition unit 101 acquires text data of I documents i (i = 1, 2,..., I) as analysis targets from the document storage unit 33. (S101). Although what kind of document group is selected as the I documents is arbitrary, for example, the following is performed.

3A to 3C are explanatory diagrams regarding an example of a method for acquiring text data of a plurality of documents to be analyzed. The items (A) and (B) described below are realized by the procedure described in the patent publication (see International Publication No. 2006/030751) filed by the present applicant. Therefore, the following description is simplified.
(A) First, processing for selecting a technology of interest from a patent gazette group of a certain company (target company) is performed. Specifically, the text data acquisition unit 101 of the processing device 1 clusters the patent documents of the target company, and obtains an evaluation value of each cluster (target company cluster) (for example, one corresponding to a factor progress information score described later). The target company cluster with the highest evaluation value is calculated and selected as the technology of interest (FIG. 3A).
(B) Next, the text data acquisition unit 101 selects a document group belonging to “selected attention technology and technology similar to the attention technology (specific technical field)” from the other patent document group including gazettes of the target company and other companies. (Own patent specific field document group) is extracted. Specifically, the text data acquisition unit 101 calculates the degree of similarity between each document of its own patent document group and the noted technology, and identifies a predetermined number of document groups having higher similarity to the noted technology. A field document group is extracted from the document storage unit 33 (FIG. 3B). Thereby, for example, after selecting an important patent group from the patent group of the target company, it is possible to analyze a similar patent group including other companies' patents.
(C) Next, the text data acquisition unit 101 obtains a document group to be analyzed by clustering the extracted patent and patent specific field document group. The group of documents to be analyzed (self-other patent cluster) is not limited to the lower cluster having a high degree of similarity between documents, but the upper cluster having a high degree of similarity between the lower clusters or an intermediate cluster in between. It may be. FIG. 3C shows an example in which 70 lower clusters, 8 middle clusters (lines), and 4 upper clusters (groups) are generated. Whether the lower cluster or the middle or upper cluster is used as the self-other patent cluster may be selected according to the purpose of analysis. As a result, for example, the patent group of a patent specific field can be classified into subdivided technical areas, and the patent groups can be analyzed for each technical area, or for each intermediate or upper cluster. . Also, the ratio of documents explained by factors in the factor analysis (cumulative contribution rate) by clustering the patent groups in the patent-specific field, which is a group of documents with high similarity, and analyzing the group of documents with higher similarity The concept or feature included in the document group can be accurately expressed.

  After acquiring the text data of each document i to be analyzed from the document storage unit 33, the text data acquisition unit 101 obtains a predetermined number J of index words j (j = 1, 2, 3,..., J). Extract. The index words to be extracted are, for example, J words having the highest importance in I documents. In order to extract the index words having the highest importance, the importance is calculated for the index words included in the document i, rearranged in descending order, and the highest J is extracted.

  The importance of the index word calculated here is the importance of the plurality of documents to be analyzed acquired by the text data acquisition unit, for example, GFIDF, or the GFIDF corrected based on the co-occurrence of index words. Is preferably used.

  GFIDF is a global frequency (GF: the total number of occurrences of the index word in the analysis target document group) and a document frequency (DF: the index word of a predetermined document group appears for a certain index word. This is a value obtained by multiplying the reciprocal of the number of documents) or the logarithm of the document frequency (IDF: inverse document frequency). A high GFIDF value is calculated for an index word that is used in a large number in the document group to be analyzed and is not frequently used in a predetermined document group different from the document group to be analyzed, and becomes an analysis target. It is evaluated as an important word representing the characteristics of the document group.

As a correction of this GFIDF based on the degree of co-occurrence between index words, there is Sky as described below. In addition, since the matter demonstrated below is implement | achieved by the procedure of the patent gazette (refer international publication 2006/048998) for which it applied by this applicant, description is simplified.
In order to obtain the Sky, first, a high-frequency word frequently appearing in the document group C to be analyzed, and the co-occurrence degree of the document unit with each index word of the document group c (w _j , w _k ) = Σ _{DεC} [DF (w _j , D) × DF (w _k , D)]
(Where w _j and w _k are index words included in the document group C, D is each document belonging to the document group C, and DF (w, D) is the document frequency of the index word w in the document D. .)
Let high frequency words that are similar to each other be a base g _h (h = 1, 2,..., B). Co (w, g) = Σ _{{w′∈g, w ′ ≠ w}} c (w, w ′)
And Here, w ′ is a high-frequency word belonging to a certain base g and other than the index word w that is a measurement target of the co-occurrence degree Co (w, g). The co-occurrence degree Co (w, g) between the index word w and the base g is the sum of the co-occurrence degrees c (w, w ′) with w for all w ′. Based on this co-occurrence degree Co (w, g), the next key (w) is calculated.
_{key (w) = 1- Π {} 1 ≦ h ≦ b} [1-Co (w, g h) / F (g h)]
Here, F (g _h ) = Σ _{wεC} Co (w, g _h ), that is, all index words w of the co-occurrence degree Co (w, g _h ) of the index word w and the base g _h Is defined as the sum of Key (w) is obtained by dividing Co (w, g _h ) by F (g _h ) to obtain a difference from 1, and multiplying all the bases g _h by taking the difference from 1, .
Sky (w) is calculated by the following equation.
Sky (w) = GF (w, C) × [IDF (w, P) + ln key (w)]
GF (w, C) is a global frequency of the index word w in the document group C to be analyzed, and IDF (w, P) is a predetermined document group P different from the document group C to be analyzed. The reverse document frequency of the index word w.
A high Sky (w) value is calculated for a word having a high GFIDF and co-occurring in the basic language of the document group C and having a high affinity with the contents of the document group C (high key (w)), and is analyzed It is evaluated as an important index word representing the characteristics of the document group C.
As described above, by extracting and analyzing only the index words representing the characteristics of the entire technical document group as important index words from multiple technical documents, the characteristics or concepts of the technical field can be understood more clearly. Is possible. On the other hand, by narrowing down the index words in advance, it becomes possible to improve the efficiency of the analysis process.

<2-2. Get document vector>
Next, after the text data acquisition unit 101 acquires text data, the document vector acquisition unit 102 generates a J-dimensional document vector having a weighting amount z of each index word j as a vector element for each document I. Generate (S102). As a result, the following I row and J column data can be obtained. Let the matrix of I rows and J columns with z as a matrix element be Z.

  Here, the weighting amount refers to a quantity given to each index word in each document from a predetermined viewpoint, and for example, TFIDF is preferably used. TFIDF is the reciprocal of index word frequency (TF: number of occurrences of the index word in a document) and document frequency (DF: number of documents of the document in which the index word appears in a predetermined document group) for a certain index word. Or it is a value calculated | required by the product with the reciprocal number (IDF: reverse document frequency) of the logarithm of document frequency. A high TFIDF value is calculated for an index word that is used in a large number in a document for which a document vector is to be calculated and is not frequently used in a predetermined document group.

<2-3. Calculation of factor loading and factor score>
Next, the factor calculation unit 103 calculates factor loadings and factor scores in factor analysis in which each document i is a subject, each index word j is an observation variable, and the document vector of each document is an answer by the subject (S103). ).
Specifically, first, the factor number of the factor k is set to K (k = 1, 2,..., K), and the factor loading for each factor k of each index word j is set to a _jk . Also, let _fik be a factor score for each factor k of each document i. Then, a factor load matrix A having a factor load a _jk as a matrix element and a factor score matrix F having a factor score f _ik as a matrix element are set as follows.

Next, let E be the residual matrix of I rows and J columns, and
Z = F × A ^t + E
However, A ^t obtains a factor loading matrix A and the factor scores matrix F by solving as a transposed matrix following A.

Regarding the factor score f _ik which is each element of the factor score matrix F and the residual e _ij which is each element of the residual matrix E, (1) the factor score is standardized to mean 0 and standard deviation 1 (2 It is assumed that the correlation between each factor score is 0, (3) the correlation between each residual is 0, and (4) the correlation between each factor score and each residual is 0. ,In general,
R = AA ^t + V
Where R is a correlation matrix between observed variables, V is a residual covariance matrix
Is known to hold. Therefore, the factor loading is obtained by the following equation.
AA ^t = R−V
Next, R−V = R ^* is set. In order to calculate this R ^* , the correlation matrix R is calculated from the values of each element z _ij of the matrix Z, and the R ^* matrix is estimated by replacing the diagonal elements of the correlation matrix with the estimated value of commonality. (Examples of commonality estimation include SMC and RMAX). Then, R * ⁼ from being a AA ^t, the R ^* matrix by calculating the factor loading matrix A based on obtaining the factor loadings (principal factor method example is as a method for obtaining the factor loadings, least squares method, There is a maximum likelihood method).

In order to find a more meaningful factor, it is desirable that the factor calculation unit 103 performs an operation of factor rotation. As a method of rotating the factor axis in the present embodiment, it is preferable to use a varimax (orthogonal rotation) method. That is, the relationship between the observed variable and the factor is obtained by rotating the factor axis while maintaining orthogonality with other factors so as to maximize the variance of the factor loading. In particular, when the document group to be analyzed is a document group having a high similarity between document vectors, this varimax method is used to reduce the amount of each factor loading, and the factor There is an advantage that the characteristics of can be clarified.
In addition to the above varimax method, as a general factor axis rotation method, for orthogonal rotation, for example, CoatMax, Ecamax, Percimax, Orthomax, Orthogonal Procrustes, etc. can be mentioned. Rotation includes Promax, Oblimin, Harris Kaiser, Oblique Procrustes, etc. The method of rotating the factor axis in the present embodiment is not limited to the above varimax, and may be appropriately selected from these rotation methods according to the embodiment.

The factor score matrix F is, for example,
F = ZR ⁻¹ A
(Where Z is standardized data here).

<2-4. Determination of attribution factor>
When the factor load matrix A and the factor score matrix F are obtained by the factor calculation unit 103, the attribute factor determination unit 104 determines the attribute factor of each index word j based on the factor load a _jk, and _sets the factor score f _ik Based on this, the attribution factor of each document i is determined (S104).
As a result, an index word and a plurality of technical documents can be attributed to each factor. Therefore, it is possible to grasp the characteristics or concept of the technical field from the index word through language information understandable to the analyst. In addition, from a plurality of technical documents, it is possible to grasp the characteristics or concepts of the technical field from the specific tendency of information included in each technical document. Although the present invention uses a factor analysis technique, it does not require an analyst's subjective or arbitrary judgment such as “interpretation of factors” in ordinary factor analysis. This is because each index word is used as an observation variable. That is, since the observation variable itself represents the content, if the observation variable belonging to the factor is determined based on the factor loading, the content of each factor can be simply indicated using the observation variable.

For example, among the factor loadings a _j1 , a _j2 ,..., A _jK for each factor of a certain index word j, if the factor loading a _jk for a certain factor k is the maximum, the attribution factor of the index word j Is the factor k. Similarly, out of the factor scores f _i1 , f _i2 ,..., F _iK for a factor of a document i, if the factor score f _ik for a factor k is the maximum, the attribution factor of the document i is determined as the factor. k. Determining the attribution factor as described above can best explain the factor by, for example, belonging to the factor most relevant to each index word or each technical document. As a result, it becomes possible to grasp the features and concepts of the technical field more clearly. In this case, one index word can belong to only one factor, and one document can belong to only one factor. On the other hand, the number of index words belonging to one factor is not necessarily one, and the number of documents belonging to one factor is not necessarily one.

Furthermore, a lower limit is set for the factor loading, and if the maximum factor a _jk of an index word j is less than the lower limit, the index word j is not attributed to any factor, and factor loading It is preferable to exclude the index word j having a low amount from the index word group indicating the content of the factor. Similarly, a lower limit value is set for the factor score, and if the maximum factor score f _ik of a document i is less than the lower limit value, the document i is not attributed to any factor, and the relationship with the factor is It is preferable to assign only high documents to factors.

  If the attribution factor of each index word is determined as described above, the index word or index word group belonging to each factor indicates the contents of each factor, that is, the concept or feature included in the I document to be analyzed. Can be considered as indicating. Since the document or document group belonging to each factor is a document or document group closely related to the factor, it can be considered to indicate which document the concept or feature extracted as each factor is particularly expressed. .

<2-5. Output>
Then, the document group analysis apparatus outputs the index word or index word group belonging to the same factor together with the document or document group data belonging to each corresponding factor by the output device 4 for each factor (S105). What is output as the data of the document or document group belonging to each factor is arbitrary, but as an example, data specifying the document or document group (publication number etc. in the case of a patent publication) is output. In addition, it is conceivable to calculate and output a factor evaluation value of the document or document group.
By calculating the factor evaluation value as described above, a relative comparison between the factors based on the factor evaluation value can be performed. As a result, it is possible to grasp the relative positional relationship between the technical elements represented by the technical literature belonging to the factors, and further classify the important technical elements from those that do not. You can also. As a preferred example of calculating the factor evaluation value, an example of calculating and outputting a factor progress information score of the document or document group will be described below.

<3. Calculation process of factor progress information score>
FIG. 2B is a flowchart for explaining a processing procedure for calculating a factor progress information score as data of a document or a document group. The process of this flowchart is as follows: (1) Document or document group belonging to each factor for which factor progress information score is to be calculated; or (2) Factor progress information score is calculated for each factor and for each applicant as described later. Is executed for each factor and each applicant's document or document group (hereinafter referred to as “score calculation target document group” in either case (1) or (2)).

<3-1. Judging the number of documents>
First, the document number determination unit 201 determines the document number N of the score calculation target document group (S201). When a published patent gazette and a patent publication gazette are issued for a certain patent application, it is desirable to count the number of documents for the patent application as two.

<3-2. Reading progress information>
Next, the progress information reading unit 202 reads the progress information as the document attribute of each document of the score calculation target document group from the document storage unit 33 of the recording device 3 or another database (S202). In this database, progress information of patent applications related to each document is recorded. As an example of the progress information to be read, for each patent application,
“Number of other company citations” (0 or positive integer),
“Number of opposition to patent or request for trial for invalidation of patent” (0 or positive integer),
“Existence request” (1 or 0),
“Patent right registration registration” (1 or 0),
"Presence / absence of accelerated examination request" (1 or 0),
"Presence or absence of appellate appeal" (1 or 0)
However, other information may be used.

<3-3. Calculation of indicators based on progress information>
Next, the index calculation unit 203 calculates a plurality of indexes based on the progress information recorded in the database as evaluation values for the score calculation target document group (S203).
Examples of this indicator include “total number of citations from other companies”, “total number of requests for opposition to patent or patent invalidation trial”, “examination request rate”, “registration assessment rate”, There are “registration rate”, “prompt request rate”, “other companies quoted ratio”, “appraisal appeal trials ratio”, “opposed objection or invalid trial request ratio”, but other indicators may be used . Each definition is as follows.
“Total number of citations from other companies” = “Total number of citations from other companies” in the document group subject to score calculation “Total number of requests for opposition to patent or request for trial for invalidity of patent”
= Total number of documents subject to score calculation for “number of oppositions to be patented or patent invalidation request” “Requested Examination Rate” = Number of Examination Requests / Number of Patent Applications “Patent Registration Rate” = Number of Patent Registrations / Number of Patent Applications “Registered assessment rate” = Number of patent registrations / Number of requests for examination “Rate of accelerated examination” = Number of requests for accelerated examination / Number of requests for examination “Ratio of citations of other companies” = “Total number of citations of other companies” / Number of patent applications Ratio of trials "= number of trials against appraisal / number of requests for examination" ratio of requests for opposition to patent or trial for invalidity of patent "
= “Total number of requests for opposition to patents or requests for invalidation of patented patents” / Number of patent registrations Note that, among these definitions, the number of requests for examination is the group of documents subject to score calculation for “whether there is a request for examination” (1 or 0) The total number of patent registrations is the total number of documents subject to score calculation for “whether or not patent rights are registered” (1 or 0). Total in the Document Group The number of trials against the appraisal may be given as the sum of the documents for which the score is to be calculated as “Presence / No Appeal Appeal” (1 or 0).

<3-4. Calculation of patent impact index>
Next, the patent impact calculation unit 204 calculates a patent impact index by giving a predetermined weight to the “number of documents” of the score calculation target document group (S204). The patent impact index refers to a document for which a score calculation target document group is to evaluate the other company's restraining power (the degree to which the rights of other companies are suppressed and the value of the company's patent is improved). For example, the “number of documents” is weighted based on the “total number of citations from other companies” and / or the “total number of requests for opposition to a patent or patent invalidation request”,
Patent Impact Index = “Number of documents” + “Total number of citations from other companies” + “Total number of requests for opposition to patent or request for trial for invalidation of patent”
Can be calculated. The weighting for the “number of documents” may be performed by addition as in the above equation or by multiplying by some other ratio.
In addition to these, the predetermined weighting includes, but is not limited to, various things such as patent profitability, patent productivity, patent utilization, and patent competitiveness.
By determining the number of technical documents as described above, it is possible to grasp the share of factors in a population composed of technical documents. In addition, by applying a predetermined weight to the number of documents, it is possible to evaluate factors that take into account economic aspects and technical competitiveness.

<3-5. Calculation of progress information index>
Next, the progress information calculation unit 205 calculates a progress information index by averaging the indices based on the progress information of the score calculation target document group (S205). The progress information index is intended to evaluate the value of patents of its own company, patent office and competitors for the score calculation target document group. For example,
Progress information index = √ {Σ _{index No = 1} ^d (index) ² / d}
Can be calculated. That is, d indicators based on the progress information, for example, “examination request rate”, “registered appraisal rate”, “patent registration rate”, “early examination request rate”, “other companies cited number ratio”, “number of appealed trial cases” The progress information index is calculated by taking the positive square root of the value calculated by dividing the sum of squares of a total of seven indicators, "Ratio" and "Ratio of objections to appeal or invalidation request" by the number of indicators d = 7. Can be calculated. Here, the above seven indicators are shown as examples of progress information indicators. However, for example, “in-house citation count ratio”, “domestic priority claim ratio”, “foreign priority claim ratio”, “packet” You may make it use "bag browsing rate" etc. By calculating the factor progress information score using the above-mentioned index, it is possible to perform factor evaluation taking into account the influence of patents that may be an obstacle to patent acquisition and technology development of other companies. In addition, it is possible to perform factor evaluations that take into account the applicant's willingness to obtain rights and examiner evaluation. Thereby, since fairness and appropriateness of factor evaluation can be ensured, as a result, the relative positional relationship between the technical elements and the result of importance can be grasped fairly and appropriately.

As a method for calculating the progress information index from the d indices, a method for obtaining the sum of the indices is also possible (simple sum method). Since a group of patents having a high value in a large number of indexes is highly evaluated, it seems reasonable to use the sum of the indexes as a historical information index. However, as the number d of indicators is increased, the specific gravity of “examination request rate”, “registration assessment rate”, “patent registration rate”, etc., which should be emphasized, decreases, and the progress information index differs from the purpose of evaluation. It can happen.
As one method for solving this problem, it is also possible to use a method in which the maximum value of each index is used as a progress information index, or a progress information index is calculated using only the top three indexes (maximum value method). However, if only the maximum value or the higher index is adopted, the other indices are not taken into account at all, and the evaluation may be one-sided.
The above-described method of taking the square root by dividing the sum of squares by the index number d can be said to be a method that combines the advantages of the simple sum method and the maximum value method. That is, by taking the sum of squares, if there is a high value index among the d indices related to a certain document group, the high value greatly affects the historical information index. Therefore, for documents that have a particularly high “request for examination rate”, “registration assessment rate”, “patent registration rate”, etc., which tend to be high (resulting in a large number of patent assessments), this is an extremely high evaluation factor. Can give points. The index other than the high value is also a progress information index that is somewhat considered.
As described above, in the present embodiment, the progress information index is calculated in consideration of all d indices. As a result, it is possible to evaluate the value of a patent from multiple angles.

<3-6. Calculation of factor progress information score>
Next, the score calculation unit 206 multiplies the patent impact index and the progress information index to calculate a factor evaluation value of the score calculation target document group (S206). By indexing the progress information in this way, for example, quantitative and objective factor evaluation can be performed.
This evaluation value is referred to as a “factor progress information score”. The calculated score is output in S105. This factor progress information score has the following properties.
When “examination request rate” = 0, the progress information index is 0 in most cases, and as a result, the factor progress information score is also 0.
The progress information index increases as the number of patents registered increases. In addition, if there are any complaints against appraisals or oppositions, they will be taken into account.
Since the patent impact index counts the number of publications, it increases as the number of patent applications increases, and further increases when patent publications are issued. The weights are weighted by “total number of other company citations” and “total number of requests for opposition to patent or requests for invalidation of patent invalid”.
This factor progress information score allows the patent document group to be evaluated from the aspect of the progress information, so that the strength of the patent that cannot be measured only by the number of patents can be known.
By calculating the factor progress information score as described above, the factor can be evaluated based on the patent publication. As a result, it becomes possible to perform highly accurate factor evaluation by utilizing examination progress information included in the patent information related to the patent publication.
The method for calculating the factor progress information score is not limited to this. For example, the individual evaluation value of each patent publication belonging to each factor is calculated, and the factor evaluation information score is obtained for each factor by summing them. You can also.

<4. Example>
Next, we introduce the analysis results of a specific document group.
A cluster analysis was performed on a patent publication and a patent publication of a certain automobile manufacturer (named company A) in Japan based on the degree of similarity to obtain a plurality of clusters. As a result of examining these multiple clusters, we decided to analyze in detail the relationship with the patents of other companies around the company, especially for the cluster related to “exhaust control of internal combustion engines”. Therefore, 2869 publications having high similarity to the publication group belonging to the cluster related to the “exhaust control of the internal combustion engine” out of about 4 million publication groups including patent publications and patent publications of the company and other companies. Groups were extracted.
The analysis result introduced below is an example in which cluster analysis is performed on the 2869 publication groups based on the similarity of the document vectors of each publication, and among these, 569 publications belonging to a specific cluster are used as the analysis target document group.

<4-1. Factor analysis processing>
In the present embodiment, among the 569 analysis target document groups, among the index words, the 77 words with the highest importance are used as observation variables, and 16 factors are extracted using the factor analysis technique. The number of factors to be extracted is not limited to this, and may be changed as appropriate according to the embodiment. Examples of methods for determining the number of factors include eigenvalues, factor contributions, interpretability, and Kaiser-Gatmann criteria.
Hereinafter, examples of the factor loading of each index word calculated by the factor calculation unit 103 and the calculation result of the factor score of each gazette are shown in the tables of FIG. 4 and FIG. Only).

The table shown in FIG. 4 is a table showing an example of the factor loading of each index word calculated for each factor of factor numbers 1 to 16. In FIG. Indicates. In FIG. 4, for convenience of explanation, not all index words are shown, and only some index words are extracted and shown.
In this example, for each index word, the factor indicating the maximum factor load (shown by shading the relevant part) shown in FIG. Specifically, for example, the index word “Kura” at the top of FIG. 4 has a factor loading of 0.96 with respect to factor 1. This index word “Kura” belongs to factor 1 because it has a larger factor loading for factor 1 than any other factors up to factors 2-16. In addition, similarly to the index word “Kura”, the index word “suck, NOx, reduction, release, spike” belongs to factor 1. However, the index words used as observation variables that did not have a factor loading equal to or greater than a predetermined threshold were not attributed to any factor.

The table shown in FIG. 5 is a table showing an example of the factor score of each gazette calculated for each factor of factor numbers 1 to 16, in which the row indicates each gazette number and the column indicates the number of each factor. ing. In FIG. 5, for the sake of convenience of explanation, a part of the publication is extracted without indicating all publications belonging to each factor, and the numbers of these publications are partially processed.
In this embodiment, as in the case of determining the attribution factor of the index word, for each gazette, the factor in which the factor score shown in FIG. 5 shows the maximum value (the corresponding part is shaded) is the gazette. The attribution factor. Specifically, for example, the publication “XX-XXX601” at the top of FIG. 5 has a factor score of 5.41 for factor 1. This publication “XX-XXX601” has a larger factor score for factor 1 than any other factors up to factors 2 to 16, and therefore belongs to factor 1. In addition, similarly to the publication of “XX-XXX601”, publications of “XX-XXX097” and “XX-XXX189” belong to factor 1. However, among the publications included in the analysis target document group, those whose maximum factor score is equal to or less than a predetermined threshold (here, 1) are not attributed to any factor.

The table shown in FIG. 6 shows index words belonging to each factor and publications belonging to each factor. FIG. 6 is a table showing examples of index words and publications belonging to each factor, but for convenience of explanation, not all publications belonging to each factor are shown, and some are extracted and shown.
As a result, even in the case of 569 analysis target document groups having a high degree of similarity, the concepts and features included in the analysis target document group can be easily understood by the index words belonging to each factor. In other words, when a document group is classified by clustering or the like, it may be difficult to grasp what is different between the classifications. However, according to the present embodiment, after extracting mutually independent concepts called factors, By assigning the most relevant index word and publication, it is possible to grasp the characteristics or concept of the technical field from the technical element indicated by the index word and the technical information included in the publication.
In addition, since it becomes possible to refer to specific gazettes for each factor as necessary, the gazette belonging to the factor to be noticed is actually read, or the factor progress information score described later is based on the gazette data belonging to each factor. More advanced analysis such as calculation is possible.

The Japanese notation differs from the notation in which a space is inserted between words such as English, and does not clearly express word breaks. For this reason, when text mining is performed on Japanese sentences, it is common to perform keyword extraction processing by applying a morphological analysis program in advance on a computer. However, the current morphological analysis program has insufficient ability to handle various expressions of Japanese natural sentences, so the words that originally make sense together are unnaturally divided, which hinders analysis. May come.
However, according to the present example, a phenomenon was observed in which the words that had been divided gathered again in the same factor. For example, the index words “Lee” and “N” belonged to Factor 4, but these were originally “Lean” (in the field of this analysis document group, “the mixing ratio of fuel and air is low”). It seems to have been an integral term. Even if the terms are divided in this way, they are gathered again in the same factor because of the high commonality of the appearance patterns of these index words, and the factor loading for each factor is also similar. Conceivable. Similarly, “Air”, “Fuel”, “Ratio”, and “Theory” of Factor 2 are considered to have been integrated terms of “theoretical air-fuel ratio”. Even in languages that use notation with spaces between words, such as English, "theoretical air-fuel ratio" may be analyzed after being divided into "sky", "fuel", "ratio", and "theory". Even in that case, if the present invention is applied, it can be expected that the same factors are collected again.

<4-2. Calculation of factor progress information score>
Next, an example in which a factor progress information score is calculated for each factor based on the progress information of the gazette belonging to each factor is shown.
Based on the progress information data of the gazette belonging to each factor, the tables shown in FIG. 7 and FIG. And the example of the result of having calculated progress information index is shown.
The examination request rate, patent registration rate, registration appraisal rate, accelerated examination request rate, other company citation number ratio, appraisal appeal ratio ratio, objection ratio ratio calculated for each factor from factor numbers 1 to 16 in FIG. And an example of a progress information index. FIG. 8 shows an example of the number of publications for each factor of factor numbers 1 to 16, the number of citations from other companies, the number of oppositions, and patent impact.

In the table shown in FIG. 9, the factor progress information score calculated by the score calculation unit 206 based on the patent impact index and the progress information index, and one publication calculated by dividing the factor progress information score by the number of documents by the score calculation unit 206 The factor progress information score (average) is shown. For reference, the eigenvalues of each factor in the factor analysis are also shown in the figure. The factor numbers in the leftmost column are common to FIGS. 4 to 8 and FIGS. 11 and 12 are arranged in the order of the factor numbers, the factor arrangement order in FIG. 9 is the descending order of the factor progress information score.
As can be seen from FIG. 9, by outputting the factor progress information score of each factor, it is possible to clearly understand which factor is the most important in the document group to be analyzed. The eigenvalue of each factor indicates how much of the data is explained, but it is understood that it is independent of the importance of the documents contained in each factor from the perspective of the factor history information score it can.

FIG. 10 shows an example in which the output device 4 outputs the factor progress information score calculated by the score calculation unit 206 in this embodiment. In FIG. 10, each circle represents each factor, the position of each circle on the vertical axis represents the factor progress information score of each factor, and the position of each circle on the horizontal axis represents the factor progress information score of each factor (the relevant factor (Average value per publication). The size of each circle indicates the number of publications belonging to each factor, and the technical terms attached to each circle indicate the index word belonging to each factor.
According to this, for example, although factor 13 (exhaust gas, NOX) has a small number of publications, it can be presumed that it is an extremely important factor from the viewpoint of progress information. Attention is paid to important technical elements in the document group to be investigated by paying attention to factors with high factor history information scores such as factor 4 (Lee, N) and factor 5 (downstream, upstream, catalyst, deterioration, diagnosis). be able to.
Specifically, for example, factor 5 (downstream, upstream, catalyst, deterioration, diagnosis) in FIG. 10 automatically determines the performance deterioration of the exhaust gas reduction device such as a catalyst system from a combination of index words of catalyst, deterioration, and diagnosis. It is presumed that this is a technical element factor that indicates a so-called advanced in-vehicle diagnosis system having a function of automatically detecting and informing the driver. Then, due to the recent exhaust gas regulations, the introduction of an advanced in-vehicle diagnostic system that monitors the malfunction of exhaust gas reduction devices such as catalysts for reducing NOx, detects it, and notifies the driver, Although requested by various councils in local governments, the importance of the technical elements is thought to be manifested in this factor 5 with meaningful index terms.

In calculating the factor progress information score, in the above example, the publication group belonging to each factor is set as the “score calculation target document group” and calculated for each factor. However, the invention is not limited to this, and further publications belonging to each factor are filed. It is good also as a "score calculation object document group" which classifies for every person and belongs to each classification for every factor and every applicant.
In the table | surface shown in FIG. 11, the calculation result of the factor progress information score for every factor and every applicant is shown.
The calculation method of the factor progress information score for each factor and for each applicant is the same as the above-described factor progress information score (for each factor) except that the score calculation target document group is changed. Since it is classified for each factor and for each applicant, the number of documents in each score calculation target document group is likely to be small. However, the examination request rate, etc. takes a value range from 0 to 1 regardless of whether or not it is classified for each applicant. Accordingly, even if the factor progress information scores for each applicant for a certain factor are summed up for all applicants, the factor progress information scores for the factors (not for each applicant) do not match.

FIG. 12 is an example in which the publication group belonging to each factor is further classified for each applicant in this embodiment, and the factor progress information score is shown for each factor and for each applicant. Specifically, the score calculation unit 206 further classifies the publication group belonging to each factor for each applicant, and calculates a factor progress information score for each factor and for each applicant. In this example, the output device 4 generates a diagram based on the calculated factor progress information score and outputs the diagram.
In FIG. 12, factors are listed in the horizontal direction in the figure, applicants are listed in the depth direction, and the height direction indicates factor progress information scores for each factor and each applicant. However, FIG. 12 shows only the results calculated for the applicant's factor progress information score with the number of filings within the top 10 in all publications with factor scores exceeding a certain level.
Specifically, for example, factor 4 and factor 5 that have been found to have a high factor progress information score in FIGS. 9 and 10 show that company A has strong power (this company A extracts a group of documents to be analyzed). The automaker (in-house) that became the source). In addition, for example, Company A is in a strong position with respect to Factor 4 and Factor 5, but is behind the other companies with respect to Factor 13, and the strengths and weaknesses between Company A and other companies, and other companies. I can grasp it.
According to this, for each factor, it is possible to grasp the rank and share of the applicants of patent publications belonging to the factor. As a result, it is possible to grasp the characteristics of a predetermined technical field from the viewpoint of a competitive state in a company that is a development subject.

The present invention is not limited to the embodiment described above, and various modifications can be made within the scope of the gist of the present invention.
In the above embodiment, the case where the technical document to be analyzed is a patent gazette is taken as an example, but the present invention is not particularly limited to this. The technical document to be analyzed may be a technical paper. In this case, the factor progress information score may be obtained by using the number of technical papers belonging to each factor, the number of citations, and the like.
In the above embodiment, as the predetermined weighting performed on the “number of documents” of the document group to be score-calculated, “total number of citations of other companies”, “total number of oppositions to be patented”, Although the case where it carries out by at least any one of "the total value of the number of claims" or all of them is taken as an example, it is not particularly limited to this. For example, it may be obtained using patent profitability, patent productivity, patent utilization, patent competitiveness, or the like.
In the above embodiment, each functional unit (text data acquisition unit 101, document vector acquisition unit 102, factor calculation unit 103, attribution factor determination unit 104, document number determination unit 201, progress information reading unit 202, Although the case where the index calculation unit 203, the patent impact calculation unit 204, the progress information calculation unit 205, and the score calculation unit 206) are realized by software is taken as an example, the present invention is not particularly limited thereto. Each functional unit of the processing device 1 may be realized by a circuit (ASIC (Application Specific Integrated Circuit) or the like) designed exclusively for executing each functional unit.
Moreover, in the said embodiment, although the processing apparatus 1 took as an example the case where the patent publications of analysis object are acquired from the memory | storage device 3, it does not specifically limit to this. For example, it is possible to communicate with an external information providing server via a network such as the Internet and acquire patent publications from the external information providing server.
In the above embodiment, the “total value of the number of oppositions to be patented” or the “total number of requests for invalidation of patents to be patented” is used as an index used for calculating the patent impact index. Is just an example. As an index used for calculating the patent impact index, both “the total value of the number of times of opposition to the patented object” and “the total value of the number of times of the patent invalidation trial request” may be used. Similarly, in the above embodiment, the “patent opposition number ratio” or “patent invalid trial request ratio” is used as an index used to calculate the progress information index, but this is merely an example. . As an index used for calculating the progress information index, both “the ratio of the number of oppositions to be patented” and “the ratio of the number of requests for invalidation of patents to be patented” may be used.

<5. Modification>
Then, the modification of embodiment mentioned above of this invention is demonstrated.
The modification of this embodiment is a factor evaluation value different from the above-described factor progress information score among the processes performed by the above-described embodiment (hereinafter, the factor evaluation value calculated in the modification is referred to as “technical element score”). Is calculated. Here, the technical element refers to each extracted factor, and is named from the technical literature included in each factor and the content of each factor represented by an index word. In the following description, a case where a patent document such as a patent publication is used as a document to be analyzed is taken as an example.

  Hereinafter, a modification of the present embodiment will be described in detail with reference to FIGS. 13 to 20. In the description of the modification of the present embodiment, the same reference numerals are used for the same configurations as in the above embodiment. Further, in the description of the present modified example, the description will focus on parts different from the above embodiment, and the description of the same configuration will be omitted.

<5-1. Configuration of Modification>
First, FIG. 13 shows a configuration of a modified example of the present embodiment. FIG. 13 is a functional block diagram of a document group analysis apparatus according to a modification of the present embodiment.

As illustrated, the document group analysis apparatus includes an input device 2, a recording device 3, an output device 4, and a processing device 100.
The input device 2, the recording device 3, and the output device 4 are the same as those in the above embodiment.
The processing device 100 classifies the patent documents for each factor according to the request of the input device 2 and uses the patent documents stored in the recording device 3 according to the above-described procedure of FIG. Further, the processing device 100 calculates a technical element score for the factor designated by the user via the input device 2.

  Specifically, the processing apparatus 100 includes a text data acquisition unit 101, a document vector acquisition unit 102, a factor calculation unit 103, an attribution factor determination unit 104, a progress information reading unit 202, a score calculation unit 2060, and a patent score calculation unit 2070. Is provided. Note that the text data acquisition unit 101, the document vector acquisition unit 102, the factor calculation unit 103, the attribution factor determination unit 104, and the progress information reading unit 202 have the same functions as those in the above embodiment, and thus description thereof is omitted here. To do.

The score calculation unit 2060 calculates a technical element score from the user via the input device 2 in a state where patent documents are classified for each factor by the attribution factor determination unit 104 and an index word is associated with each factor. Accept the request.
When the score calculation unit 2060 receives a calculation request for a technical element score, the score calculation unit 2060 calculates a technical element score by using a “patent score (PS)” indicating an evaluation value for each patent document belonging to a calculation target factor. The “patent score (PS)” is calculated in advance by the patent score calculation unit 2070 shown below.

  The patent score calculation unit 2070, for each patent document, progress information of the patent document (information such as presence / absence of priority claim, number of citations in examination of other patent applications) and content information (number of claims, , The “patent score (PS)” that evaluates the patent document is calculated. Then, for each piece of information (gazette number) for identifying a patent document, the patent score calculation unit 2070 “patent score (PS)” of the patent document and “abandonment information” indicating whether the patent is abandoned. (Information including whether or not rejection is confirmed is included) is generated (hereinafter referred to as “PS information”).

Next, the hardware configuration of the processing apparatus 100 will be described. As in the above-described embodiment, the processing device 100 includes a CPU (Central Processing Unit), a memory, an I / F that exchanges data with external devices (such as the input device 2, the recording device 3, and the output device 4). It is realized by a computer provided.
Then, each functional unit of the processing device 100 (text data acquisition unit 101, document vector acquisition unit 102, factor calculation unit 103, attribution factor determination unit 104, progress information reading unit 202, score calculation unit 2060, and patent score calculation unit 2070). ) Shall be realized by software.

  Specifically, in the memory of the processing apparatus 100, each functional unit (text data acquisition unit 101, document vector acquisition unit 102, factor calculation unit 103, attribution factor determination unit 104, progress information reading unit 202, score calculation unit 2060 is stored. , And a program for realizing the tent score calculation unit 2070) is stored. And each function part of the processing apparatus 100 shall be implement | achieved when said CPU (Central Processing Unit) runs said program memorize | stored in memory.

<5-2. Calculation processing in modification>
Next, a technical element score calculation process according to a modified example different from the above-described embodiment will be described.
FIG. 14 is a flowchart illustrating a procedure of a technical element score calculation process according to a modification of the embodiment of this invention.
The processing flow shown below outputs data of documents or document groups belonging to each factor after the determination of attribution factors (S104) of each document in the factor extraction processing of FIG. 2A described above (S104). S105). Further, “information in which patent documents belonging to factors are associated with each factor (information shown in FIG. 6)” obtained by the factor extraction processing in FIG. 2A is stored in a predetermined area of the memory of the processing device 100. It is assumed that
Further, it is assumed that the patent score (PS) for each patent document belonging to each factor is calculated by the patent score calculation unit 2070 before performing the processing of FIG. In the memory (or storage device 3) of the processing device 100, for each piece of information (gazette number) identifying the patent document, the “patent score (PS)” of the patent document and whether the patent is waived. It is assumed that information (hereinafter referred to as “PS information”) associated with “abandonment information (including information indicating whether rejection has been confirmed)” indicating whether or not is stored is stored. The procedure for calculating the patent score (PS) will be described with reference to FIGS.

Specifically, the score calculation unit 2060 receives a request for a technical element score calculation process from the user via the input device 2 (S2010). In addition, when requesting the calculation process of the technical element score, the user also designates a category to be calculated.
For example, a factor obtained by the factor extraction process in FIG. 2A may be specified as the classification target. In this case, a technical element score is calculated for each factor.
Further, for example, as a classification subject to calculation, patent gazettes belonging to each factor may be classified for each applicant, and a classification for each factor and each applicant may be designated. In this case, a technical element score is calculated for each factor and for each applicant.
In the following, a case where a request for calculating a technical element score of a certain factor is received is taken as an example.

Next, the score calculation unit 2060 acquires a patent score (PS) of a patent document belonging to a factor that is a calculation target of the technical element score received in S2010 (S2020).
Specifically, the score calculation unit 2060 uses “information in which a patent document is associated with each factor (information shown in FIG. 6)” and “PS information” stored in the memory of the processing device 100. The “patent score (PS)” and “abandonment information” of the patent document belonging to the factor to be calculated are acquired.

  Next, the score calculation unit 2060 uses the “patent score (PS)” and “waiver information” of the patent document belonging to the acquired factor to be calculated, and for each patent score (PS) that has not been waived. The standard value is obtained (S2030).

Specifically, the score calculation unit 2060 refers to the “waiver information” and, among the patent documents belonging to the designated factor, patent patents (including applications pending with the Patent Office) that have not been surrendered. Specify the score (PS).
The score calculation unit 2060 obtains a standard value for each identified patent score (PS) in a population (for example, a patent document that has not been surrendered in the analysis target document group subjected to factor extraction processing). More specifically, the score calculation unit 2060 obtains a standard value for each identified patent score (PS) using the following (Equation 1) and the identified patent score (PS).

In (Equation 1) shown below, it is assumed that there are “m” patent scores (PS) of patent documents that have not been waived in the population, and a subscript i is added to the patent score (PS), and “PSi ( 1 ≦ i ≦ m (m is an integer of 1 or more)) ”.
In (Expression 1), the standard value of “patent score PSj” of each patent document j belonging to a specific factor is obtained from the PSi of m patent documents.

  Next, the score calculation unit 2060 obtains the total value of the standard values of the patent score PSj greater than or equal to the threshold among the standard values of the patent scores PSj of the patent documents belonging to the specific factor obtained in S2030, and calculates the total value. The “technical element score” of the factor is set (S2040). In this step, the score calculation unit 206 obtains the maximum value among the standard values of the patent scores PSj of the patent documents belonging to the specific factor obtained in S2030.

Specifically, the score calculation unit 206 uses the following (Equation 2) and the standard value of the patent score (PSj) obtained in S2030 to calculate the “technical element score” for the factor specified by the user. calculate. Further, the score calculation unit 206 selects the maximum (MAX) standard value from the standard values of the patent scores PSj obtained in S2030, and sets the selected standard value as the maximum value in the factor.
In (Expression 2), among the standard values of each patent score PSj obtained in S2030, it is assumed that the number of standard values of the patent score PSj equal to or greater than the threshold is “n” in the factor. In (Expression 2), as an example of the threshold value PSstd, the average of the standard values of each patent score PSi obtained in S2030 (0 according to [Expression 1]) is used.

Then, when the technical element score is calculated by the score calculation unit 2060, the process proceeds to S105 (output) in FIG.
In the flow of FIG. 14, the technical element score for one factor is calculated, but this is merely an example. When a request for calculating the technical element score of a plurality of factors is received, the processing of S2020 to S2040 is performed for each factor, and the technical element score and the maximum value are obtained for each factor.

In S105 of FIG. 2A, the output element 4 outputs the technical element score obtained in S2040. Alternatively, the output device 4 outputs the maximum value of the factor together with the technical element score.
The score calculation unit 2060 obtains a technical element score for each factor and each applicant, and generates information indicating a graph indicating the technical element score for each factor and each applicant as shown in FIG. The output device 4 may output the data. In this case, the technical element score and the maximum value may be shown together for each factor and for each applicant.

As described above, in the present modification, the technical element score is calculated using the patent score (PSi) of the patent document that has not been waived. The reason for this is as follows. For example, when a company tries to evaluate patents in each technical field, the number of patent documents classified into a certain technical field (factor) is very large, but many of them are abandoned (or rejected). Suppose that the application has been finalized. In such a case, if an application that has already been abandoned (or an application for which refusal has been finalized) is included in the evaluation of a patent in that technical field, the technical field that does not hold many patent rights will be highly evaluated. Therefore, proper analysis is not possible.
Therefore, in this modification, the technical element score is calculated using the patent score (PSi) of a patent document that has not been abandoned so as to improve the accuracy of the score.

  Further, in this modification, when calculating the standard value of the patent score (PSi), a general standard value is multiplied by a coefficient instead of a mere standard value ((Equation 1) is multiplied by 10). ing). This is for facilitating the discrimination between the obtained standard values. In (Equation 1), it is 10 times, but it is only an example.

In this modification, only the standard value of the patent score PSi exceeding the threshold is used for calculating the technical element score. This is to alleviate the influence of the number of patent documents received by the value of the technical element score.
For example, suppose that a technical element score is obtained for each applicant and for each factor, and the technical tendency for each applicant is analyzed by comparing the obtained technical element scores. In this case, if the threshold value is not taken into consideration as in the present modification, the value of the technical element score of the applicant having a large number of applications tends to be too high, and there is a possibility that a highly accurate analysis cannot be performed.
Certainly, if the patents of a specific technical field are extracted without excess or deficiency and used as the analysis target document group (population), the number of applications for each applicant and each factor itself is also a sufficiently significant value. Can think. However, if the analysis target document group (population) is extracted by an arbitrary method that is not so, the number of applications itself varies depending on the characteristics of the industry to which each applicant belongs. If the number of applications for each factor is limited, there is a possibility that a highly accurate analysis cannot be performed.
In addition, when the focus is on selecting important elements from a group of documents to be analyzed (population) including a huge number of patents, the “individual importance” is more than the “large number of patents with low individual importance”. In some cases, it is preferable to focus on those that include “high patents”.
Therefore, in the present modification, only the standard value of the patent score PSi that is equal to or higher than a predetermined value is used, and a high technical element score is given only to factors including important patents that are higher than the predetermined value. In order to improve the accuracy of the technical element score.
In particular, for example, when the patent score is standardized so that the average becomes 0 and the standard value equal to or higher than the average (0) is totaled as the technical element score, not only the patent score value below the average can be discarded, Even if there are a large number of patent scores near the average, the influence on the value of the technical element score is small, and if there is something that is higher than the average, the value of the technical element score is greatly affected. Therefore, it is possible to further reduce the influence of the number of cases included in the technical elements and accurately extract the technical elements including the patents with high importance.

FIG. 15 is a diagram showing the distribution of the technical element score of the modified example and the factor progress information score of the example in relation to the number of publications. Specifically, a factor progress information score is calculated for a plurality of factors extracted from a group of documents to be analyzed and standardized to an average of 0 and variance 1, and a technical factor score (average 0 and variance 1 is standardized for the same factor. ), And the normalized factor progress information score and technical element score are plotted on the vertical axis, and the number of publications for each factor is plotted on the horizontal axis.
As shown in the figure, the factor progress information score has a distribution close to a straight line showing a direct proportional relationship with the number of publications, and is greatly influenced by the number of publications. On the other hand, although the technical element score is not completely irrelevant to the number of publications, it is distributed in a region considerably away from a straight line showing a direct proportional relationship, which shows that the influence of the number of publications is mitigated.

FIG. 16 shows that after determining the attribution factor of each document (S104) in the factor extraction process of FIG. 2A described above, the publication group belonging to each factor is further classified for each applicant, In this example, the technical element score of this modification is calculated for each person. A specific calculation method is as follows.
A cluster analysis was performed on a Japanese patent publication and a patent publication of a house equipment manufacturer (company a) based on the degree of similarity to obtain a plurality of clusters. As a result of examining these multiple clusters, we decided to analyze in detail the relationship with the patents of other companies around the company, especially for the “composite structure” cluster. Therefore, from about 4 million gazette groups including patent publications and patent publication gazettes of the company and other companies, about 3000 gazette groups having a high similarity to the gazette group belonging to the cluster related to the “composite structure”. Extracted.
A cluster analysis was performed on the approximately 3000 publication groups based on the similarity of the document vectors of each publication, and among these, 323 publications belonging to a specific cluster were extracted as factors to be analyzed. The score calculation unit 2060 further classifies the publication group belonging to each factor for each applicant, and calculates the technical element score for each factor and for each applicant. Then, the output device 4 generates a diagram based on the calculated technical element score, and outputs the diagram.
In FIG. 16, factors are listed in the horizontal direction in the figure, applicants are listed in the depth direction in the order of the number of applications, and the height direction indicates the technical element score for each factor and each applicant. However, FIG. 16 shows only the results calculated for the applicant's technical element score with the number of applications within the top 10 in all publications with factor scores exceeding a certain level.

In this modification, since the technical score is calculated excluding publications with a patent score below the average, the factors that account for the majority of publications below the average, or the factors that all publications are below the average For, the technical element score is close to 0 or 0. Therefore, the contrast between factors becomes clear, and as a result, the order and evaluation between factors can be easily grasped visually.
From the output result of FIG. 16, it can be seen that company g, which has a small number of applications in the analysis target document group, has a patent group comparable to company a and company b, which have a large number of applications, in a certain technical element. In addition, it is clear at a glance where the strengths and weaknesses are, even in the case of company a, which is the top in the number of applications, such as the existence of technical elements that lag behind other companies. In this way, it is possible to clearly grasp the strengths and weaknesses of each company.

In this modification, the average of the population is used as the threshold, but the present invention is not particularly limited to this. For example, an average of the standard values of the patent score PSi in the patent group of the specific applicant and other threshold values determined by other users may be set in the processing apparatus 100.
In this modification, the standard value of the patent score PSi is used, but the present invention is not particularly limited to this. For example, the influence of the number of cases can be mitigated even when only non-standardized patent scores PSi are added that are greater than or equal to a predetermined value.

Moreover, according to this modification, when the technical element score is presented to the user, the highest standard value of the patent score (PSj) of the patent document classified as the factor can be presented. As a result, the user can grasp which technical element (factor) the highly evaluated patent is included in. Accordingly, even if the evaluation value as a whole of the technical elements (factors) is low, the user can grasp the technical elements (factors) including the highly evaluated patent.
For example, assume that a company seeks to evaluate a patent for each technical field and obtains a technical element score for each factor of the company (applicant). In this case, by presenting the highest value of each factor, it becomes possible to grasp which technical field of the company has strong patents.

<6. About Patent Score (PS)>
Next, the patent score (PS) used for calculating the technical element score in the above modification will be described with reference to FIGS.
In the following description, the patent score (PS) calculation process is performed by the patent score calculation unit 2070 of the processing apparatus 100. However, the present invention is not limited to this.
Another computer having a CPU (Central Processing Unit), a memory, and the like may perform the patent score calculation process. In this case, a program (PS calculation program) for realizing the function of the patent score calculation unit 2070 is stored in another computer. Then, the CPU of another computer executes the “PS calculation program”, thereby calculating the patent score PS and generating the above-described PS information. The processing apparatus 100 acquires PS information generated by another computer and stores it in the memory. When another computer performs the patent score calculation process, the processing apparatus 100 need not include the patent score calculation unit 2060.

<6-1. Data structure>
First, the data structure used for calculating the patent score PS will be described.
The storage device 3 stores patent data (electronic data indicating a patent publication) and patent attribute information. The electronic data indicating the patent publication includes at least the patent data ID (gazette number, etc.), the application date, and the bibliographic information such as the IPC code.
The patent attribute information includes progress information 300 of the patent document (information such as whether priority is claimed or the number of citations in examination of other patent applications), and content information 400 (number of claims, Information such as the number of specifications). Hereinafter, the data structure of the progress information 300 and the content information 400 will be described.

First, an example of the data structure of the progress information 300 is shown in FIG.
FIG. 17 is a diagram schematically illustrating an example of the data configuration of the progress information used in the modification of the present embodiment.
As shown in the figure, the progress information 300 includes a field 301 for registering “patent data ID (gazette number, etc.)”, a field 302 for registering “the number of days elapsed from the filing date”, and “examination request date”. A field 303 for registering "elapsed days since", a field 304 for registering "elapsed days since registration date", a field 305 for registering information indicating whether "divisional application" exists, A field 306 for registering information indicating the presence / absence of “early examination”, a field 307 for registering information indicating the presence / absence of “trial decision of appeal”, and information indicating the presence / absence of “opposition maintenance decision” are registered. A field 308 for registering, a field 309 for registering information indicating the presence / absence of “invalidation trial decision”, and information indicating the presence / absence of “priority claim” A field 310 for registering information indicating the presence / absence of “PCT application”, a field 312 for registering information indicating the presence / absence of “packaging browsing”, and “number of times cited” And a field 313 for registering the information to be shown constitutes one record. The progress information 300 includes a plurality of records.

  Here, “elapsed days from application”, “elapsed days from examination request”, and “elapsed days from registration date” are information relating to the period of the corresponding patent data. “Elapsed days from application” is the application date, “Elapsed days from examination request” is the application examination request date, and “Elapsed days from registration date” is the evaluation date (calculation of patent score). The number of days elapsed up to a predetermined date close to the evaluation date is calculated and stored in the storage device 3. “Elapsed days from examination request” for a patent application that has not yet been requested for examination of application is NULL, and “elapsed days from registration date” for a patent application that has not yet been set and registered is NULL.

  Among the progress information 300, “divisional application”, “accelerated examination”, “approval appeal decision”, “opposition maintenance decision”, “invalidity decision maintenance decision”, “packaging browsing”, “priority” This is information indicating the presence or absence of a predetermined action on the data. "Divisional application" is whether the divisional application has been filed based on the patent application, "Rapid examination" is whether the patent application has been expedited, and Whether an appeal against a decision to reject the application has been requested and whether a patent trial decision has been made in that trial, whether or not the opposition maintenance decision has been made, whether or not a patent opposition has been made and a maintenance decision has been made on the patent, The “invalidation trial maintenance decision” is whether the patent invalidation trial has been requested for the patent, and whether the appeal has been rejected in the trial, “priority” is the priority claim based on the previous patent application etc. Whether or not the patent application is an international application based on the Patent Cooperation Treaty and whether or not it is a domestic application. Based on whether it is, respectively given 1 for example if the predetermined action has been performed, if not been given a 0, for example.

Next, the data structure of the content information 400 is shown in FIG.
FIG. 18 is a diagram schematically illustrating an example of a data configuration of content information used in a modification of the present embodiment.

As shown in the figure, the content information 400 includes a field 401 for registering “patent data ID (gazette number, etc.)”, a field 402 for registering “number of claims” of the patent data, and “claim One record is composed of a field 403 for registering the “average number of characters” and a field 404 for registering the “number of specifications” of the patent data. The content information 400 includes a plurality of records.
Here, the “number of claims” is information indicating the number of claims of the patent application, and the “average number of characters of the claim” is the average number of characters (or the number of words) per claim of the patent application. It is information to show. The “number of specification pages” is information indicating the number of specification pages or publication pages of the patent application. Such information is extracted from published patent gazettes and other patent data of each patent application.

<6-2. Patent score calculation processing>
Next, description will be made with reference to FIG. FIG. 19 is a flowchart illustrating a procedure of a patent score calculation process according to a modification of the present embodiment.

As shown in FIG. 19, the patent score calculation unit 2070 receives an input of an IPC code from the user, and acquires patent data (electronic data indicating a patent publication) (S400).
Specifically, when receiving an IPC code input from a user, the patent score calculation unit 207 accesses the storage device 3 and acquires patent data classified into the IPC code. The patent data includes bibliographic information such as the filing date information and priority date information of the patent application (only when priority is claimed).

Next, the patent score calculation unit 2070 uses the application date information or the priority date information among the obtained bibliographic information of the patent data, and converts the patent data for each predetermined period (in the modification of this embodiment, for each application year). , For each year to which the priority date belongs, etc.) (S500).
Next, the patent score calculation unit 2070 calculates an evaluation value of each patent data (S600). Details of this processing will be described with reference to FIG.

FIG. 20 is a flowchart showing details of processing for calculating an evaluation value of each patent data according to a modification of the present embodiment.
The patent score calculation unit 2070 acquires the progress information 300 and the content information 400 for the patent data belonging to the group generated by the classification of S210 (S610). Specifically, the patent score calculation unit 2070 uses the patent ID (gazette number, etc.) included in the bibliographic information of the acquired patent data to store the progress information 300 and the content information 400 stored in the storage device 3. The progress information 300 and the content information 400 associated with the patent ID of the acquired patent data are acquired from the inside.
Here, in FIG. 20, it is assumed that the acquired one group consists of J patent data, and a subscript j (j = 1, 2,..., J) is used to distinguish each of the J cases. .
When J patent data is acquired, using the J patent data progress information 300 and content information 400, “total value for J of the evaluation item corresponding data” used in S6302 to S6304, which will be described later, etc. Is obtained in advance.

  Next, the variable j is set to 1 (S620), and the evaluation raw score of the patent data j is calculated as follows.

  First, information registered in each field of the progress information 300 is used as an evaluation item, and I evaluation items i (i = 1, 2,..., I) are evaluated in advance for each evaluation item. A calculation method is selected (S6301).

  There are the following three evaluation point calculation methods in the modification of the present embodiment. That is, for information registered in the fields 305, 306, 307, 308, 309, 310, 311 and 312, S6302 [Presence / absence type] is selected as information indicating the presence / absence of a predetermined action on the patent data. For fields 302, 303, and 304, S6303 [time decay type] is selected as information related to the period of the patent data. In the field 313, S6304 [number-of-times] is selected as information indicating the number of times the patent data is cited.

  When the evaluation score calculation method is selected, the evaluation score of the patent data j is calculated for each of the I evaluation items i (S6302, S6303, S6304).

<6-2-1. Calculation of evaluation score for presence / absence type>
For the evaluation item i for which S6302 [presence / absence type] is selected, the evaluation score is calculated by the following [Equation 3].

  Here, the “existence / non-existence data of the evaluation item i” arranged in the numerator is, for example, “1” if the divisional application has been filed as described above, and “0” if it has not been made.

In the denominator, the positive square root of the in-group total value of the above “evaluation item i presence / absence data” is arranged. Therefore, the denominator is large when there are many patent data corresponding to the evaluation items in the group, and the denominator is small when there are only a few patent data corresponding to the evaluation items in the group. Patents with fewer evaluation items (such as “Invalidation Trial Maintenance Decision”) than patents with a higher number of evaluation items (such as “Bag Viewing”) will be maintained after patent registration (In general, a high maintenance rate is considered to indicate a high economic value commensurate with the maintenance cost (patent fee)), and thus each evaluation item is automatically weighted. In addition, since it is tabulated in groups for each predetermined period, for example, older patents have more progress information added, and new patents that have just been published often do not yet have progress information added. It can alleviate the tendency for new patents to be given low ratings for that reason.
The attribute information of the patent data is useful for relative evaluation within the analysis target population, but proper evaluation cannot be performed if the patent applications or patent rights in the analysis target population are treated equally. According to the present embodiment, the analysis object population including patent applications or patent rights at different periods is classified by classifying the analysis object population into groups for each period and using the value obtained for each classified group as a denominator. Appropriate relative assessment is possible within the population.
For example, in a certain technical field, the former value is often higher between one value in a simultaneous group with few patent applications and one value in a simultaneous group with many patent applications. On the other hand, for example, a patent application that has passed several years is more likely to be given progress information, such as a request for browsing, than a patent application that has just been published. It is an error to underestimate a patent application that has just been made. For example, if only a few of the patent applications in the same period group have been requested to be browsed, the patent application that has received the request for browsing is a patent application with a particularly high degree of attention and should be highly evaluated. On the other hand, if there are a large number of requests for inspection among patent applications in the same period group, the patent application that received the request for inspection should be highly evaluated just because it was requested for inspection. is not.
According to the present embodiment, the value obtained using the patent attribute information of each patent data belonging to each group and the value obtained using the patent attribute information of each patent data belonging to the group are determined for each group. The evaluation score is calculated by multiplying the sum of the values by the value of the decreasing function. According to this structure, the value which considered the relative positioning of each patent data in each group can be calculated | required as an evaluation value. As a result, the lower the total value of the numerical information based on the progress information in the simultaneous group, the higher the weight, and conversely the lower the higher the total value, the lower the weight, so that Appropriate evaluation is possible.

<6-2-2. Calculation of evaluation points for time decay type>
For the evaluation item i for which S6303 [Time decay type] is selected, the evaluation score is calculated by the following [Equation 4].

  “Exp (− (Min (elapsed time, year limit)) / year limit)” placed in the numerator here is the “number of days elapsed since the request for examination”. ”Or“ year ”, which is the value obtained by dividing the smaller value by“ year ”and multiplying by −1 and the power of the Napier number e. The “year” is the maximum number of years from the filing date until the expiration of the patent right (20 years under the current Japanese law). The same formula is used for “elapsed days from registration date”, and “year” is the maximum number of years from the filing date to the expiration of the patent term (20 years under the current Japanese law). The same formula is used for the “elapsed days from the filing date”, but the “year” is the number of years from the filing date to the application examination request deadline (3 years in the current Japanese law). According to this, while the elapsed time is short, the value of the numerator is a value close to Exp (0) = 1. However, when the elapsed time ≧ years, the value of Exp (−1) = 1 / e. To fall. The advantage of using an exponential function is that a depreciation effect on the value can be introduced and that the evaluation value distribution can be eliminated and a smooth distribution can be achieved. “Elapsed days from request for examination”, “Elapsed days from application date”, and “Elapsed days from registration date” are basic evaluation items applicable to many patents. Tying can be avoided.

  The denominator has the same formula as the above S6302 [Presence / absence type], but the “days since examination request” is, for example, 1 if an application examination request is made for the patent application, and if not, for example 0 Are summed within the group to obtain a positive square root. For the “elapsed days from the date of registration”, the denominator is a value obtained by adding a value of 1 within the group by taking the positive square root by adding 1 if the patent application has been registered for patent right setting and not being registered. . Since all patent data falls under “Elapsed days since filing”, the value of the denominator is equal to the positive square root of the number of patent data in the group, assuming that the evaluation data of the relevant evaluation item is 1. . In any case, the denominator is large when there are many patent data corresponding to the evaluation items in the group, and the denominator is small when there are only a few patent data corresponding to the evaluation items in the group. As described above, “Elapsed days from request for examination”, “Elapsed days from application date”, and “Elapsed days from registration date” are basic evaluation items applicable to many patents. Tends to be small.

The evaluation score calculated in S6303 [time decay type] is further corrected by content information.
In the following, the content information 400 shown in FIG. 18 is used.
When the evaluation is based only on the progress information, there is a possibility that the patent application or the patent right shortly after the publication of the application or the registration of the patent right does not have the progress information expected to be granted in the future and the evaluation cannot be performed correctly. Therefore, in order to correct this, content information is added to the evaluation based on the progress information. However, the content information tends not to have a high correlation with the maintenance rate as the progress information, and if the content information is inadvertently added, the accuracy of the evaluation may decrease.
Therefore, in order to keep the influence of the content information small in the evaluation of a patent with sufficient progress information and to effectively reflect the content information in the evaluation of a patent with insufficient progress information, this S223C [time decay type] Only the evaluation score calculated in (5) is multiplied by the correction coefficient based on the content information.
As described above, according to the present embodiment, regardless of whether the application is old or new, it is possible to add the content information of each patent data to the information related to the period having characteristics that are easily given to any patent data. As a result, it is possible to perform appropriate evaluation even for patent data consisting of a new application to which little progress information is given.

Specifically, for each evaluation point in [Equation 4],
a ₁ × a ₂ × a ₃
here,
a ₁ = 2 ^1/3 (when the average number of characters per claim is below average) or 2 ^-1/3 (when the average number of characters per claim is above average)
a ₂ = 2 ^1/3 (when the total number of pages is above average) or 2 ^-1/3 (when the total number of pages is below average)
a ₃ = 2 ^1/3 (when the number of claims is within an average value ± 1 standard deviation) or 2 ^−1/3 (when the number of claims is outside the above range)
Multiply By setting the maximum values of a ₁ , a ₂ , and a ₃ to 2 ^1/3 , the correction is limited to a ₁ × a ₂ × a ₃ as the maximum value. In the above embodiment, the value of a ₁ × a ₂ × a ₃ is set to 2 at the maximum.

<6-2-3. Calculation of evaluation points for the frequency type>
For the evaluation item i for which S6304 [number-of-times] is selected, an evaluation score is calculated by the following [Equation 5].

Here, “f (quotation) × log (n _j +1)” arranged in the numerator is the weight of the logarithm of the value obtained by adding 1 to the “cited count n _j ” for the “cited count”. Quoting). According to the verification by the present inventors, it has been found that the maintenance rate of the patent right changes depending on the number of citations as well as the presence or absence of citations. Since the increase gradually shows a tendency to peak, the logarithm is taken.

In the denominator, the positive square root of the total value in the group of “f (quotation) × log (n _j +1)” is arranged. Accordingly, the denominator is large when there are many patent data cited in other applications in the group, and the denominator is small when there are only a few patent data cited in other applications in the group.

具体的な重みとしては、他社引用の場合のｆ（引用_other）と、自社引用の場合のｆ（引用_self）との比を、１：２とした。In the numerator and denominator of [Formula 5], an arbitrary positive number can be used as the weight f (quotation), but the number of times cited in other patent applications (number of times other companies cited) n _jother and other patents of the company It is distinguished from the number of times cited in the application (in-house citation number) n _jself, and a different weight is _assigned to each logarithm. In this case, instead of the above [Equation 5], the following [Equation 6] is used.

As a specific weight, the ratio of f (quoting _other ) in the case of _other company citations and f (quoting _self ) in the case of company citations was set to 1: 2.

The number of times cited is highly correlated with the value of a patent. Furthermore, according to the verification by the present inventors, the number of times cited in the examination of patent applications of other companies (citation of other companies) and the number of times cited (in-house quotation) in examinations of other patent applications of the company are the latter. Was found to be significantly higher in correlation with patent value. The invention cited in the examination of other patent applications of the company is presumed to be due to the fact that it is often the basic invention that is the core in the implementation technology of the company. And while recognizing that the company has already applied for such a basic invention, there is a high possibility that the company has applied for the improved technology and built a strong patent portfolio.
According to this embodiment, it is possible to appropriately evaluate a patent application or a patent right by considering the number of citations separately from other company citations and company citations, and reflecting the latter number more largely in the evaluation value.

この式に示されるように、評価素点は、Ｉ個の評価点の二乗和の正の平方根、又は０となる。評価素点が０となるのは、審査請求期限までに出願審査請求しなかった場合、出願を取下げ又は放棄した場合、拒絶査定が確定した場合、その他特許出願が失効した場合と、異議申立による取消決定や無効審判による無効審決が確定した場合、特許権を放棄した場合、特許権の存続期間が満了した場合、その他の特許権が消滅した場合である。これらの情報も各特許データの経過情報から読み取り、該当する場合は評価素点を０とする。
上述のようにＳ６３０３〔時間減衰型〕で算出された評価点に対しては、内容情報による補正を行う。具体的には、「審査請求からの経過日数」、「出願日からの経過日数」、「登録日からの経過日数」に基づき上述の［数４］で算出された評価点にそれぞれ上述のａ_１×ａ_２×ａ_３を乗算した上で、［数７］に従い二乗和の平方根をとる。<6-2-4. Calculation of evaluation score>
When the evaluation score of the patent data j is calculated for all the evaluation items i (i = 1, 2,..., I), the evaluation score of the patent data j is calculated based on the evaluation score of the following [Expression 7]. ] (S640).

As shown in this equation, the evaluation raw score is a positive square root of the sum of squares of I evaluation points, or 0. The evaluation score is 0 because the application request is not requested by the deadline for requesting examination, the application is withdrawn or abandoned, the decision of refusal is finalized, other patent applications have expired, The case where the decision of revocation or the trial for invalidation by the trial for invalidation is finalized, the patent right is abandoned, the duration of the patent right expires, or the other patent right is extinguished. These pieces of information are also read from the progress information of each patent data, and the evaluation raw score is set to 0 when applicable.
As described above, the evaluation score calculated in S6303 [time decay type] is corrected by the content information. Specifically, each of the evaluation points calculated in the above [Equation 4] based on “the number of days elapsed since the request for examination”, “the number of days elapsed since the application date”, and “the number of days elapsed since the registration date” _After multiplying by ₁ × a ₂ × a ₃ , the square root of the sum of squares is taken according to [Equation 7].

As a method of calculating an evaluation raw score from an evaluation point i based on a plurality of evaluation items, there is a method of calculating a sum of each evaluation point i (simple sum method). However, according to this calculation method, since the evaluation of a patent to which a lot of historical information having a correlation with the patent maintenance rate (economic value) is given is calculated high, the sum of the evaluation points i should be used as an evaluation raw score. Is reasonable at first glance, but the evaluation score of a patent that has been granted a lot of historical information that does not have a high correlation with the maintenance rate (a lot of low evaluation points are added) has a very high correlation with the maintenance rate Care should be taken because it may exceed the evaluation score of a patent to which a small amount of progress information is granted.
As one method for solving this problem, there is a method in which the maximum value among the evaluation points i is used as an evaluation raw score (maximum value method). However, according to this calculation method, especially when investigating the correlation between certain historical information and the maintenance rate of a group of patents, when investigating the correlation regardless of what other historical information is given, maintaining a certain patent Since the rate is expected to be best expressed by the maintenance rate of the historical information having the highest maintenance rate, it is reasonable to use the maximum value of the evaluation point i as an evaluation raw score. If the maximum value is the same in the two patents, no superiority or inferiority is given. Furthermore, when the maximum value method is used, it is not possible to carry out an evaluation that takes into account the viewpoints of three different entities of the applicant, the JPO, and competitors, and only the viewpoints of any one of those entities The viewpoint of the remaining subject cannot be reflected in the evaluation of patent data.
The above-described method for taking the square root of the sum of squares can be said to be a method that combines the advantages of the simple sum method and the maximum value method. That is, by taking the square root of the sum of squares, when there is a high evaluation point i in I evaluation items i related to a certain patent data j, the high evaluation point i greatly affects the evaluation raw score. The evaluation points other than the evaluation item having a high evaluation point i are also evaluation raw points that are somewhat considered. Therefore, a high evaluation score is given to patent data j that corresponds to multiple items such as “early examination”, “opposition to maintain opposition”, and “invalidation maintenance decision” that tend to be high. be able to.
Thus, in this modification, patent evaluation is performed in consideration of all evaluation points calculated according to the type of patent attribute information (S630, S640). As a result, it is possible to evaluate the value of patent data from multiple aspects.

<6-2-5. Calculation of evaluation value>
When the evaluation raw score is calculated, its logarithm is calculated and used as the evaluation value of the patent data j (S650).
The evaluation value calculated based on the progress information or content information is given a high value for a few patent applications or patent rights whose unique progress or content can be read, but for many other patent applications or patent rights. Is often given a low value. Accordingly, looking at the distribution of the number of evaluation values, the number of patent applications or patent rights with high evaluation values is a few and sparse distribution, and the number of patent applications or patent rights with low evaluation values is a dense distribution.
In such a case, the average value (arithmetic average value) is greatly influenced by a small number of patent applications or patent rights with high evaluation values, so care must be taken when evaluating by comparison with such average values. It becomes. In addition, for example, when comparing two patent applications or patent rights that have obtained high evaluation values, even if it appears that there is a large difference in evaluation values, it may not be a significant difference in practice. is there.

Next, it is determined whether or not evaluation values have been calculated for all patent data j (S660). If not calculated (S660: NO), the process proceeds to S670, the variable j is set to j + 1, and the process returns to S630. The evaluation value is calculated for the following patent data.
When the evaluation values are calculated for all the patent data j (S660: YES), the evaluation value calculation processing for the patent data belonging to the group ends.
As described above, in the present embodiment, a plurality of patent data having different characteristics are evaluated in consideration of the characteristics for each technical field and each filing time. As a result, the value of patent data can be more appropriately evaluated.

The evaluation value calculation processing from S610 to S670 is executed for all the groups t obtained by classifying the patent data acquired in S400 in S500.
When the evaluation values are calculated for all the groups t, the processing returns to FIG. 19, and the deviation value in the analysis target population acquired in S400 is calculated as the patent score PS based on the evaluation values (S700). This deviation value also enables relative comparison of patent data between different technical fields that are difficult to compare (comparison with a population to be analyzed separately selected by different IPCs in S400). is there.

And in the modification of this embodiment, since the technical element score is calculated based on the patent score PS obtained by the above procedure, there are the following advantages compared to the above embodiment. .
Specifically, in the above modification, the patent score PS that is the basis of the technical element score takes into account the weight according to the type of progress information. Since the technical element score is obtained using the patent score PS, a higher accuracy score is calculated in the modification.
According to the patent score of this modification, the analysis target population is classified into groups for each period, and the values obtained for each classified group are used as denominators to include patent applications or patent rights at different periods. Appropriate relative evaluation is possible within the analysis population.
Therefore, it is possible to reduce the possibility that a high evaluation value is calculated for the technical element score of a factor in which many patent data whose applications are old are classified.

Claims (14)

Text data acquisition means for acquiring a plurality of technical documents represented by text data;
For each acquired technical document, a weighting amount calculating means for obtaining a weighting amount of each index word,
Using each obtained technical document as a subject and using the obtained weighting amount of each index word, perform factor analysis using each index word as an observation variable, and for each index word, factor loading for each factor And calculating means for calculating a factor score for each factor for each of the technical documents,
An attribute factor determining means for determining an attribution factor of each index word using the factor loading of each index word, and determining an attribution factor of each technical document using a factor score of each technical document,
A document group analysis apparatus comprising: an output unit that outputs an index word or a group of index words belonging to the same factor together with data of technical documents or technical document groups belonging to the corresponding factors, for each factor. The document group analysis apparatus according to claim 1,
The attribution factor determination means includes
For each index word, using the calculated factor loading, the factor with the largest factor loading is selected, the selected factor is specified as the attribution factor of the index word, and each technical document A document group analysis apparatus characterized in that, using the calculated factor score, the factor having the maximum factor score is selected and the selected factor is specified as the attribution factor of the technical document. The document group analysis apparatus according to claim 1 or 2,
The frequency of occurrence of index words included in the plurality of technical documents is obtained, the importance level of each index word is calculated using the frequency of appearance, and a predetermined number of index words having the highest importance level are calculated using the calculated importance level. It further comprises an important index word extraction means for extracting,
The document group analysis apparatus characterized in that the weighting amount calculating means obtains a weighting amount of a predetermined number of index words with higher importance as a weighting amount of each index word. The document group analysis apparatus according to any one of claims 1 to 3,
For each of the factors, a factor evaluation value calculation means for calculating a factor evaluation value indicating a technical evaluation of the factor,
The output unit outputs a factor evaluation value of the technical document or technical document group as data of the technical document or technical document group. The document group analysis apparatus according to any one of claims 1 to 3,
The technical literature is a patent document including a patent publication and a patent publication,
Progress information acquisition means for acquiring progress information of each patent document;
For each of the factors, using the progress information of the technical literature or technical literature group belonging to the factor, a factor evaluation value calculating means for calculating a factor evaluation value indicating a technical evaluation of the factor, and
The document group analysis apparatus characterized in that the output means outputs the factor evaluation value as data of the technical document or technical document group. The document group analysis apparatus according to claim 5,
Provided with document number determination means for determining the number of technical documents or technical documents belonging to each factor,
The factor evaluation value calculation means includes
For each of the factors, a first index is calculated by giving a predetermined weight to the number of documents in the technical document or technical document group belonging to the factor. 2. A document group analyzing apparatus that calculates two indexes and calculates a factor evaluation value of the factor using the calculated first index and second index. The document group analysis apparatus according to claim 6,
The progress information includes the number of citations from other companies, the number of oppositions to be patented, the number of requests for a patent invalidation trial, the existence of a request for examination, and the presence or absence of registration of patent right setting,
The first index is weighted using at least one of the total number of citations of other companies, the total number of patent oppositions, and the total number of patent invalidation requests. Value
The second index, which is an index of the progress information, is the total number of citations from other companies, the total number of oppositions to be patented, the total number of requests for patent invalidation trials, the examination request rate, and the registration assessment rate. A document group analyzing apparatus characterized in that it is a value obtained by indexing at least one of them. The document group analysis apparatus according to any one of claims 5 to 7,
The document group analysis apparatus characterized in that the output means calculates the factor evaluation value for each factor and for each applicant. A data analysis method performed by an information processing apparatus,
The information processing apparatus includes:
Obtaining a plurality of technical documents represented by text data;
Obtaining a weighting amount of each index word for each acquired technical document;
Using each obtained technical document as a subject and using the obtained weighting amount of each index word, perform factor analysis using each index word as an observation variable, and for each index word, factor loading for each factor And for each of the technical documents, calculating a factor score for each factor;
Determining the attribution factor of each index word using the factor loading of each index word, and determining the attribution factor of each technical document using the factor score of each technical document. Data analysis method. A program for causing an information processing device to execute data analysis processing,
The program is
A process of obtaining a plurality of technical documents represented by text data;
For each technical document acquired, a process for obtaining a weighting amount of each index word;
Using each obtained technical document as a subject and using the obtained weighting amount of each index word, perform factor analysis using each index word as an observation variable, and for each index word, factor loading for each factor And calculating the factor score for each factor for each of the technical documents,
Determine the attribution factor of each index word using the factor loading of each index word, and cause the information processing device to execute the process of determining the attribution factor of each technical document using the factor score of each technical document A program characterized by
The document group analysis apparatus according to any one of claims 1 to 3,
The technical literature is a patent document including a patent publication and a patent publication,
For each acquired patent document, means for acquiring a patent score that individually evaluates the value of the patent document;
Document group analysis comprising factor evaluation value calculation means for calculating a factor evaluation value indicating a technical evaluation of the factor for each factor using the patent score of a patent document belonging to the factor apparatus. The document group analysis apparatus according to claim 11,
The factor evaluation value calculation means includes
For each factor, a patent score that is equal to or higher than a predetermined threshold is selected from among the patent scores of patent documents belonging to the factor, and a value obtained by tabulating the selected patent scores is calculated as the factor evaluation value. Document group analyzer. The document group analysis apparatus according to claim 12,
The document score analyzing apparatus, wherein the patent score is a value standardized in a document group of a population including a factor for which the factor evaluation value is calculated. The document group analysis apparatus according to any one of claims 11 to 13,
The patent score means that the patent documents are classified into groups for each technical field and for each predetermined period, and for each classified group, the progress information of the patent documents belonging to the group is used, and for each patent document, A document group analyzing apparatus characterized by being a calculated value.

Images