JP6781123B2 - Data processing equipment, data processing method and data processing program - Google Patents

Description

The present invention relates to a data processing apparatus, a data processing method and a data processing program.

Conventionally, extraction of related words related to a target word has to be done manually by an expert, and it takes time, so there is a problem that it is difficult to keep updating the related word dictionary with the latest information. Therefore, it is expected to automate the extraction of such related term terms.

Therefore, in the field of natural language processing, a technique for automatically extracting related terms related to a target word from document data has been proposed. For example, as a conventional method, when a word that appears in the same sentence as a word in a document or around it is co-occurred with the word, and the word is related to the word. , There is a way to extract this word as a related word. Further, a method of extracting related words from a document in a field that matches the target field even when related words do not necessarily co-occur in one document (see, for example, Patent Document 1) and a method of high importance. A method has been proposed in which words are selected, the degree of relevance between important words is determined, and related words that are closely related are extracted (see, for example, Patent Document 2).

Japanese Unexamined Patent Publication No. 2004-361992 Japanese Unexamined Patent Publication No. 2003-167894

However, in the conventional method, all co-occurrence words are used as related word candidates. Therefore, in the conventional method, even if the related words are limited by the number of occurrences and the frequency of appearance, the conceptual words having a high degree of abstraction are included, and the words having low relevance or words having no relevance are included. There was a problem that many related words were mixed and the accuracy of extracting related words was low.

The present invention has been made in view of the above, and an object of the present invention is to provide a data processing apparatus, a data processing method, and a data processing program capable of extracting related words from document data with high accuracy.

In order to solve the above-mentioned problems and achieve the object, the data processing apparatus according to the present invention extracts related word candidates related to the target word from the document data by coexistence of words, and the related words of each target word. An acquisition unit that acquires a candidate group, a counting unit that counts the number of occurrences in a plurality of related word candidate groups for each related word candidate included in a plurality of related word candidate groups, and an appearance counted by the counting unit. It is characterized by having a related word determination unit that excludes related word candidates whose number is equal to or greater than a predetermined threshold from a plurality of related word candidate groups and determines that the remaining related word candidates are related words of the target word. And.

According to the present invention, related words can be extracted from document data with high accuracy.

FIG. 1 is a diagram schematically showing an example of a configuration of a data processing device according to an embodiment. FIG. 2 is a diagram illustrating a processing flow of the data processing apparatus shown in FIG. FIG. 3 is a diagram illustrating a processing flow of the data processing apparatus shown in FIG. FIG. 4 is a flowchart showing a processing procedure of the data processing method according to the embodiment. FIG. 5 is a flowchart showing a processing procedure of the related word candidate appearance number counting process shown in FIG. FIG. 6 is a diagram showing an example of a computer in which a data processing device is realized by executing a program.

Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings. The present invention is not limited to this embodiment. Further, in the description of the drawings, the same parts are indicated by the same reference numerals.

[Embodiment]
Embodiments of the present invention will be described. In the embodiment of the present invention, it is assumed that the object is digitized text document data. Then, in the present embodiment, among the related word candidates in the related word candidate group of each target word extracted from the document data by co-occurrence of words, the number of occurrences in the plurality of related word candidate groups is equal to or higher than the threshold value. Exclude certain related word candidates and use the remaining related word candidates as related words of the target word. The target word is a word that is the target of the related word extraction process, and the related word is a word that is extracted from the document data by co-occurrence of words as a word related to the target word.

[Data processing device configuration]
First, the configuration of the data processing device according to the embodiment will be described. FIG. 1 is a diagram schematically showing an example of a configuration of a data processing device according to an embodiment. As shown in FIG. 1, the data processing device 1 includes an input unit 11, an output unit 12, a communication unit 13, a control unit 14, and a storage unit 15.

The input unit 11 is an input interface that receives various operations from the operator of the data processing device 1. For example, the input unit 11 is composed of an input device such as a touch panel, a voice input device, and a keyboard and a mouse.

The communication unit 13 is a communication interface for transmitting and receiving various information to and from other devices connected via a network or the like. The communication unit 13 is realized by a NIC (Network Interface Card) or the like, and communicates between another device and the control unit 14 (described later) via a telecommunication line such as a LAN (Local Area Network) or the Internet. For example, the communication unit 13 receives the data of the electronic document file via the network and outputs the data to the control unit 14. Further, the communication unit 13 outputs information indicating technical terms generated by the control unit 14 to an external device via the network.

The output unit 12 is realized by, for example, a display device such as a liquid crystal display, a printing device such as a printer, an information communication device, or the like, and outputs information or the like indicating a related word of the target word generated by the control unit 14.

The control unit 14 controls the entire data processing device 1. The control unit 14 is, for example, an electronic circuit such as a CPU (Central Processing Unit) or an MPU (Micro Processing Unit), or an integrated circuit such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array). In addition, the control unit 14 has an internal memory for storing programs and control data that define various processing procedures, and executes each process using the internal memory. In addition, the control unit 14 functions as various processing units by operating various programs. The control unit 14 has a related word candidate group acquisition unit 141 (acquisition unit), a related word candidate appearance number counting unit 142 (counting unit), and a related word determination unit 143.

The related word candidate group acquisition unit 141 extracts related word candidates related to the target word by co-occurrence of words from the document data, and acquires the related word candidate group for each target word. The related word candidate group acquisition unit 141 targets nouns and compound nouns extracted from the document data by using morphological analysis, and uses morphological analysis from the part of the document data including the target word for each target word. Using the extracted nouns and compound nouns as related word candidates, a related word candidate group that summarizes related word candidates for each target word is acquired. Specifically, the related word candidate group acquisition unit 141 extracts nouns and compound nouns from the document data to be processed by using morphological analysis, and creates a target word list for each of these extracted words. To do. Then, the related word candidate group acquisition unit 141 extracts, for each word in the target word list, a portion including this word from the document data. The related word candidate group acquisition unit 141 extracts nouns and compound nouns from the extracted part by using morphological analysis, and uses the extracted words as related word candidates to collect related word candidates for each target word. get.

The related word candidate appearance count unit 142 counts the number of occurrences in the plurality of related word candidate groups for each related word candidate included in the plurality of related word candidate groups acquired by the related word candidate group acquisition unit 141. ..

The related word determination unit 143 excludes the related word candidates whose counted number of occurrences is equal to or greater than a predetermined threshold value from the plurality of related word candidate groups, and determines that the remaining related word candidates are related words of the target word. The related word determination unit 143 has a related word candidate extraction unit 144, a related word candidate exclusion unit 145, and a related word data storage unit 146.

The related word candidate extraction unit 144 extracts related word candidates whose number of occurrences counted by the related word candidate appearance number counting unit 142 is equal to or greater than a predetermined threshold value from a plurality of related word candidate groups. At this time, the related word candidate extraction unit 144 performs the extraction process using the threshold value set according to the document data to be processed.

Here, this threshold value is changed according to the document data to be processed. For example, the threshold value is appropriately set according to the amount of document data, the data content of the document data, the creation period of the document data, and the like. Further, the threshold value may be set by simulation based on the processing contents accumulated in the past document data processing by the data processing device 1, the field of the document data to be processed, the amount of data, the creation period, and the like. For example, the threshold is half the total number of target words. Of course, the threshold value may be one-third of the total number of target words or one-fourth of the total number of target words. Further, the threshold value is not limited to the total number of target words, but may be set according to the total number of related word candidate groups and the number of each related word candidate included in the related word candidate group. For example, the input unit 11 changes the threshold value by receiving the instruction information instructing the setting or change of the threshold value. Alternatively, the related word candidate extraction unit 144 may change the threshold value according to a predetermined rule.

The related word candidate exclusion unit 145 excludes the related word candidates extracted by the related word candidate extraction unit 144 from the plurality of related word candidate groups. The related word data storage unit 146 determines that the related word candidate remaining after the exclusion of the related word candidate exclusion unit 145 is a related word of the target word, and stores the related word data 154 (described later) in the storage unit 15.

The storage unit 15 is a storage device for an HDD (Hard Disk Drive), an SSD (Solid State Drive), an optical disk, or the like. The storage unit 15 may be a semiconductor memory in which data such as a RAM (Random Access Memory), a flash memory, and an NVSRAM (Non Volatile Static Random Access Memory) can be rewritten. The storage unit 15 stores an OS (Operating System) and various programs executed by the data processing device 1. Further, the storage unit 15 stores various information used in executing the program. The storage unit 15 stores document data 151, count data 152, threshold data 153, and related word data 154.

The document data 151 is digitized text document data, and includes a document file to be processed by the data processing device 1. The count data 152 is data in which each count number counted by the related word candidate appearance number counting unit 142 is associated with each related word candidate. The threshold data 153 is data indicating a threshold value that can be changed according to the document file to be processed. Further, the related word data 154 is data in which each related word is associated with each target word.

[Data processing flow]
Next, the processing flow in the data processing apparatus 1 will be described in detail. 2 and 3 are diagrams illustrating a processing flow of the data processing apparatus shown in FIG.

First, with reference to FIG. 2, the process up to the acquisition of the related word candidate group of the target word will be described. As shown in FIG. 2, the related word candidate group acquisition unit 141 extracts nouns and compound nouns from the electronic file document 151-1 to be processed by morphological analysis, and targets each of these extracted words as the target word. to create a word list _{L 1} (see (1) Figure 2). For example, the target word list L _1, as a target word, "Tokoro division", "Emergency", shown in the order of "transmission". Further, the target word list L _1, a description hereinafter as included are N target word.

The related term candidate group acquiring unit 141, a text sentence of the electronic file documents 151-1, sentence line file F ₁ of the type shown sentence in _line, paragraph line file F ₂ of the type shown the paragraph on one line _or, , Transforms into a one-bar, one-line file with one bar per line (see (2) in FIG. 2).

Then, description will be given of a case where to get one second of related words candidates of "Tokoro division" of the target word of the target word list L _1. In this case, the related word candidate group acquiring unit 141, a line containing the "Tokoro split", sentence line files F _1, paragraph line file F _2, and is extracted from a bar line file ((3 2) see ). For example, related words candidate group acquiring unit 141, to extract the sentence to be "bug to fail the transmission of the emergency call at Tokoro split has occurred." From the sentence line file F _1. Thus, the related word candidate group acquiring unit 141, a plurality extracts the line containing "Tokoro split" as shown in the data P _1.

Then, the related word candidate group acquisition unit 141 extracts nouns and compound nouns from the line including the extracted "place division" by morphological analysis, and removes the duplication as the related word candidate of "place division". (See (4) in FIG. 2). That is, the related word candidate group acquisition unit 141 sets a collection of related word candidates including "emergency call", "send", "failure", "bug", and "occurrence" as related word candidates of the target word "place division". obtaining as a group _{G 1.}

The related term candidate group acquiring unit 141, the second of the target word of the target word list L ₁ proceeds to processing for "Emergency", Similarly, "Emergency" in FIG. 2 (3), (4 ) Is repeated (see (5) in FIG. 2). As a result, the related word candidate group acquisition unit 141 acquires the related word candidate group of the target word “emergency call”. Thus, for each target word of the target word list L _1, (3) in FIG. 2, by repeating the processing described in (4), to acquire the related word candidates for each target language. Related term candidate group acquiring unit 141 acquires the related word candidates _G 1 ~G _N for 1~N pieces of target language respectively.

Next, the processing of the related word candidate appearance number counting unit 142 and the related word determination unit 143 will be described with reference to FIG. As shown in FIG. 3, the related term candidate group acquiring unit 141 acquires the N pieces of object words respectively associated word candidates G ₁ ~G _N (see (6) in FIG. 3), the related word candidate appearance count parts 142, for each target word, counting the number of occurrences in all related term candidates _G 1 ~G _N of each related term candidate (see (7) in FIG. 3).

The related term candidate extraction unit 144, a certain number of all related words candidates G ₁ ~G _N (e.g., half of all target word number) extracts a related word candidate appearing in common to the above (in FIG. 3 (See (8)). Here, in the upper frame 3, all related words a fixed number of candidates G ₁ ~G _N (e.g., total target word number half) above appearing word (related term candidate) is star to the right It is marked.

For example, related term candidate extraction unit 144, the related term candidate group G ₁ of "Tokoro resolution" total target word number more than half, in all relevant word candidate group G ₁ ~G _N, related words appearing in common "transmission" as a candidate, "failure", "bugs" extracts "generation" (central word reference group G _{1 'in} the frame of FIG. 3). The related term candidate extraction unit 144, from the related term candidate group G ₂ of "Emergency", the total target word number more than half, in all relevant word candidate group G ₁ ~G _N, related words appearing in common "function" as a candidate, "send", to extract the "generation" (central word group G _{2 'references} the framework of FIG. 3). The related term candidate extraction unit 144, from the related term candidate group G ₃ of the "transmission", the total target word number more than half, in all relevant word candidate group G ₁ ~G _N, related word candidates commonly occurring as for extracting the "bugs", "failure", "function" (central word reference group G _{3 'in} the frame of FIG. 3).

Subsequently, the related word candidate excluding unit 145, extracted by the related term candidate extraction unit 144, a predetermined number or more, in all the relevant word candidates G ₁ ~G _N, excluding the emerging words in common, related word data The storage unit 146 sets the remaining related word candidates as related words of each target word (see (9) in FIG. 3). Then, the related word data storage unit 146 associates the related word with the target word and stores it in the storage unit 15.

For example, the related word data storage unit 146 is a remaining associated word candidate of the associated word candidate group G ₁ of "Tokoro split" to "emergency call", the bottom of the related terms K _{1 (FIG.} 3 "Tokoro split" It is stored in the storage unit 15 as (see in the frame). Related word data storage unit 146 is a remaining associated word candidate of the associated word candidate group G ₂ of "Emergency", "Tokoro division", the "number notification", related terms K _{2 (figure} "Emergency" It is stored in the storage unit 15 as (see in the lower frame of 3). Further, the related word data storage unit 146 is a remaining associated word candidate of the associated word candidate group G ₃ of "transmit,""receive," and "data", the related word K _{3 (FIG.} 3 of the "transmission" It is stored in the storage unit 15 as (see in the lower frame). As a result, the related words in the lower frame of FIG. 3 are stored in the storage unit 15 as related word data in a state of being associated with each target word.

[Processing procedure of data processing method]
Next, with reference to FIG. 4, the processing procedure of the data processing method by the data processing apparatus 1 shown in FIG. 1 will be described. FIG. 4 is a flowchart showing a processing procedure of the data processing method according to the embodiment.

First, as shown in FIG. 4, when the control unit 14 reads the document data to be processed, the related word candidate group acquisition unit 141 extracts the related word candidates of the target word from the document data by co-occurrence of words. Then, a related word candidate group is acquired for each target word (step S1). The related word candidate appearance count unit 142 performs a related word candidate appearance count process for counting the number of occurrences in the plurality of related word candidate groups for each related word candidate included in the plurality of related word candidate groups (). Step S2).

Subsequently, in the related word determination unit 143, the related word candidate extraction unit 144 refers to the threshold value according to the document data (step S3), and the number of occurrences counted by the related word candidate appearance number counting unit 142 is referred to. Related word candidates that are equal to or greater than the threshold value are extracted from a plurality of related word candidate groups (step S4). Subsequently, the related word candidate exclusion unit 145 excludes the related word candidates extracted by the related word candidate extraction unit 144 from the plurality of related word candidate groups (step S5).

Then, the related word data storage unit 146 determines that the related word candidate remaining after the exclusion of the related word candidate exclusion unit 145 is a related word of the target word, and stores it in the storage unit 15 as the related word data 154 (step). S6).

[Processing procedure for counting the number of related word candidates]
Next, the processing procedure of the related word candidate appearance number counting process will be described with reference to FIG. FIG. 5 is a flowchart showing a processing procedure of the related word candidate appearance number counting process shown in FIG.

As shown in FIG. 5, the related word candidate appearance number counting unit 142 initializes n, which is the identification number of the target word, and sets n = 1 (step S11). The related term candidate appearance counting section 142 sets the related word candidates for the first counted among the related word candidates associated word candidates G _n of subject words n (step S12). If n = 1, the related term candidate appearance number counting unit 142, first, the first target word of the target word list (hereinafter, the target word 1.) The first related related word candidate group G ₁ of Set word candidates as counting targets.

The related term candidate appearance counting section 142, the related word candidates counted to count the number of occurrences in each related term candidates G ₁ ~G _N (step S13). Related term candidate appearance number counting unit 142, for all the relevant word candidates associated word candidates Gn, determines whether or not count the number of occurrences in each related term candidates G ₁ ~G _N (step S14).

Related term candidate appearance number counting unit 142, for all the relevant word candidates associated word candidates Gn, if it is determined that no count the number of occurrences in each related term candidates G ₁ ~G _N (step S14: No ), The related word candidate to be counted next in the related word candidate group _Gn for the nth target word in the target word list is set (step S15). For example, related word candidate appearing counting section 142, when the count for the target language 1 is finished, the second related word candidates associated word candidates G _1, is set as the count target. Thus, the related word candidate appearing counting section 142, for all relevant word candidates associated word candidate group G ₁ of the target word 1, repeat steps S13~ step S15.

On the other hand, related word candidate appearing counting section 142, for all the relevant word candidates associated word candidates Gn, if it is determined that the counted number of occurrences in each related term candidates G ₁ ~G _N (step S14: Yes ), Determine whether or not the target words 1 to N have been counted (step S16). When the related word candidate appearance number counting unit 142 determines that the target words 1 to N have not been counted (step S16: No), n = n + 1 is set for the identification number n of the target word (step S17). Specifically, the related word candidate appearing counting section 142, when it is completed to count for all relevant word candidates associated word candidate group G ₁ of the target word 1, the second object of the target word list L ₁ Proceed to the word (hereinafter referred to as the target word 2). The related term candidate appearance counting section 142, the related word candidates associated word candidate group G ₂ of the target word 2, successively, counts the number of occurrences in each related term candidates G ₁ ~G _N.

When the related word candidate appearance count unit 142 determines that the target words 1 to N have been counted (step S16: Yes), each related word candidate group G is used for all the related word candidates of all the target words 1 to N. since counting the number of occurrences in ₁ ~G _N, and terminates the related term candidate appearance number counting process.

[Effect of Embodiment]
As described above, the data processing device 1 according to the present embodiment extracts the related word candidates of the target word from the document data by co-occurrence of words, and acquires the related word candidate group of each target word. Then, the data processing device 1 counts the number of occurrences in the plurality of related word candidate groups for each related word candidate included in the plurality of related word candidate groups, and the counted number of occurrences is equal to or higher than a predetermined threshold value. A certain related word candidate is excluded from a plurality of related word candidate groups, and the remaining related word candidate is used as a related word of the target word.

Here, in the conventional technique, related words related to the target word are automatically extracted from the document data by co-occurrence of words, but all the co-occurrence words are used as related word candidates, so that the relevance is high. Low words or unrelated words were included in the related words, and the accuracy of extracting the related word group was low.

Since these words with low relevance or words with no relevance appear in large numbers and are considered to be words with a relatively high degree of abstraction, they often co-occur in common to different target words. Therefore, in the present embodiment, the words that appear in common in the related word candidate group acquired for each target word are words that are related to any target word but have a weak connection with the target word. We focused on the fact that there are many cases, that is, the words are often less relevant to the target word.

Then, in the present embodiment, only words that are strongly related to the target word are narrowed down by removing words that appear in common in these related word candidate groups. In other words, in the present embodiment, words that appear in common in a certain number or more of the related word candidate groups of each target word are extracted, these words are excluded, and only words that are highly related to the target word are selected. By narrowing down, related words are extracted. Therefore, according to the present embodiment, words having low relevance or words having no relevance can be appropriately excluded as compared with the conventional data processing method, so that the related words can be extracted with high accuracy. Can be extracted.

[System configuration, etc.]
Each component of each of the illustrated devices is a functional concept and does not necessarily have to be physically configured as shown in the figure. That is, the specific form of distribution / integration of each device is not limited to the one shown in the figure, and all or part of the device is functionally or physically distributed in arbitrary units according to various loads and usage conditions. It can be integrated and configured. Further, each processing function performed by each device may be realized by a CPU and a program analyzed and executed by the CPU, or may be realized as hardware by wired logic.

Further, among the processes described in the present embodiment, all or a part of the processes described as being automatically performed can be manually performed, or the processes described as being manually performed can be performed. All or part of it can be done automatically by a known method. In addition, the processing procedure, control procedure, specific name, and information including various data and parameters shown in the above document and drawings can be arbitrarily changed unless otherwise specified.

[program]
FIG. 6 is a diagram showing an example of a computer in which the data processing device 1 is realized by executing a program. The computer 1000 has, for example, a memory 1010 and a CPU 1020. The computer 1000 also has a hard disk drive interface 1030, a disk drive interface 1040, a serial port interface 1050, a video adapter 1060, and a network interface 1070. Each of these parts is connected by a bus 1080.

The memory 1010 includes a ROM (Read Only Memory) 1011 and a RAM 1012. The ROM 1011 stores, for example, a boot program such as a BIOS (Basic Input Output System). The hard disk drive interface 1030 is connected to the hard disk drive 1090. The disk drive interface 1040 is connected to the disk drive 1100. For example, a removable storage medium such as a magnetic disk or an optical disk is inserted into the disk drive 1100. The serial port interface 1050 is connected to, for example, a mouse 1110 and a keyboard 1120. The video adapter 1060 is connected to, for example, the display 1130.

The hard disk drive 1090 stores, for example, the OS 1091, the application program 1092, the program module 1093, and the program data 1094. That is, the program that defines each process of the data processing device 1 is implemented as a program module 1093 in which a code that can be executed by a computer is described. The program module 1093 is stored in, for example, the hard disk drive 1090. For example, a program module 1093 for executing processing similar to the functional configuration in the data processing device 1 is stored in the hard disk drive 1090. The hard disk drive 1090 may be replaced by an SSD.

Further, the setting data used in the processing of the above-described embodiment is stored as program data 1094 in, for example, a memory 1010 or a hard disk drive 1090. Then, the CPU 1020 reads the program module 1093 and the program data 1094 stored in the memory 1010 and the hard disk drive 1090 into the RAM 1012 and executes them as needed.

The program module 1093 and the program data 1094 are not limited to those stored in the hard disk drive 1090, but may be stored in, for example, a removable storage medium and read by the CPU 1020 via the disk drive 1100 or the like. Alternatively, the program module 1093 and the program data 1094 may be stored in another computer connected via a network (LAN, WAN (Wide Area Network), etc.). Then, the program module 1093 and the program data 1094 may be read by the CPU 1020 from another computer via the network interface 1070.

Although the embodiment to which the invention made by the present inventor is applied has been described above, the present invention is not limited by the description and the drawings which form a part of the disclosure of the present invention according to the present embodiment. That is, all other embodiments, examples, operational techniques, and the like made by those skilled in the art based on the present embodiment are included in the scope of the present invention.

1 Data processing device 11 Input unit 12 Output unit 13 Communication unit 14 Control unit 15 Storage unit 141 Related word candidate group acquisition unit 142 Related word candidate appearance count unit 143 Related word judgment unit 144 Related word candidate extraction unit 145 Related word candidate exclusion Part 146 Related word data storage part 151 Document data 152 Count data 153 Threshold data 154 Related word data

Claims (5)

An acquisition unit that extracts related word candidates related to the target word by co-occurrence of words from the document data and acquires a related word candidate group for each of the target words.
For each related word candidate included in the plurality of related word candidate groups, a counting unit that counts the number of occurrences in the plurality of related word candidate groups, and a counting unit.
Related word candidates whose number of occurrences counted by the counting unit is equal to or greater than a predetermined threshold value are excluded from the plurality of related word candidate groups, and the remaining related word candidates are determined to be related words of the target word. Word judgment part and
A data processing device characterized by having. The acquisition unit uses morphological analysis to extract nomenclature and compound nomenclature extracted from the document data using morphological analysis, and extracts each target word from the portion of the document data including the target word. The data processing apparatus according to claim 1, wherein a related word candidate group in which related word candidates are grouped for each target word is acquired by using a noun and a compound nomenclature as related word candidates. The data processing apparatus according to claim 1 or 2, wherein the predetermined threshold value is changed according to document data. A data processing method performed by a data processing device.
The process of extracting related word candidates of the target word from the document data by co-occurrence of words and acquiring the related word candidate group of each of the target words, and
A step of counting the number of occurrences in the plurality of related word candidate groups for each related word candidate included in the plurality of related word candidate groups, and
The related word candidates whose counted number of occurrences is equal to or greater than a predetermined threshold are excluded from the plurality of related word candidate groups, and the related word candidates remaining in the related word candidate group are the targets corresponding to the related word candidate group. The process of determining that the word is related to the word,
A data processing method characterized by including. A step of extracting related word candidates related to the target word by co-occurrence of words from the document data and acquiring a related word candidate group for each of the target words.
For each related word candidate included in the plurality of related word candidate groups, a step of counting the number of occurrences in the plurality of related word candidate groups, and
The related word candidates whose counted number of occurrences is equal to or greater than a predetermined threshold are excluded from the plurality of related word candidate groups, and the related word candidates remaining in the related word candidate group are the targets corresponding to the related word candidate group. Steps to determine that the word is related to the word,
A data processing program that allows a computer to run.

Images