JP2018077670A - Analysis device, analysis method and analysis program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an analysis device for appropriately generating dependency between each value of items in different documents.SOLUTION: An analysis device stores classification information in which a calculation condition that defines a calculation range of similarity and classification conditions related to a plurality of feature information for values of items corresponding to a type are associated with each other, and stores a plurality of first feature information relating to a value of a first item described in a first document and a plurality of second feature information relating to a value of a second item described in a second document. A processor of the analysis device refers to the classification conditions to specify the type of the first item based on the plurality of first feature information, specify the type of the value of the second item based on the plurality of second feature information, and specify the calculation condition corresponding to the type of the first item; calculates the similarity between the value of the first item and the value of the second item according to the calculation condition when the type of the first item and the type of the second item are the same; and generates dependency information indicating dependency between the value of the first item and the value of the second item using the similarity.SELECTED DRAWING: Figure 6

Description

  The present invention relates to an analysis apparatus, an analysis method, and an analysis program for analyzing data.

  In the business processing of information systems such as data editing processing, there is a demand for a technique for efficiently analyzing and tracing the dependency relationship between design documents for the purpose of grasping the dependency relationship between design specifications and investigating the influence.

  Patent Document 1 discloses a technique for extracting related information indicating the relationship between requirement element information and design element information. In Patent Document 1, “a plurality of requirement element information included in the requirement definition document and requirement element item information that is item information included in the requirement element information are acquired from the requirement definition document, and the basic design is obtained from the basic design document. A plurality of design element information included in a document and design element item information which is item information included in the design element information, and the design element item information included in the design element information among the acquired plurality of design element information. A combination of design element information whose intersection set is not an empty set is extracted, a union of the extracted design element item information is generated, one of the acquired requirement element information is extracted, and the extracted requirement element information has When the requirement element item information is compared with the generated union, and the requirement element item information is included in the union at a predetermined ratio or more, the requirement element item information A set of the design element information and response requirements element information corresponding to the union, there is a description that extracts "as related information indicating the association between the requirements element information and said design element information.

  Patent Document 2 discloses a technique for detecting a relationship between items in a plurality of tables. According to Patent Document 2, “for each item in the table, an item evaluation step for calculating an item score that is a value based on the number of appearances for each value in the item is associated with a different table from the plurality of tables. Select as a source table and a related destination table, select a related source item that is an item in the related source table and a related destination item that is an item in the related destination table, and select the related source item and the related destination item. A relation generation step for generating a combination as a relation, and for each relation generated by the relation generation step, the number of appearances for each value in the relation source item in the relation destination item, the item score of the relation source item, and the relation There is a description that "the computer executes a related evaluation step of calculating a related score that is a value based on the item score of the previous item".

JP 2013-008237 A JP 2008-27072 A

  In Patent Document 1, the relevance of a set of a plurality of items is investigated from the ratio of matching items. Therefore, in the case where there is a large number of notation fluctuations in the items, for example, and they do not match, the technique of Patent Document 1 cannot extract the relationship between the items. For example, in the technique of Patent Document 1, whether the requirement element item of the requirement element item information is included in “classification”, “customer code”, “representative name”, etc., which are the union of design element item information. To check. At this time, even if “customer code” exists in the requirement element item and the item is synonymous with “customer code” in the design element item information, the technique of Patent Document 1 traces between them due to notation fluctuation. Can not do it.

  Further, the technique of Patent Literature 2 cannot detect a match when there is a notation of a value held in the table and the values do not match. Therefore, when the rate of notation fluctuation is high and the decrease in the related score is significant, the technique of Patent Literature 2 may not be able to detect the relationship between the tables. Furthermore, even if the association between tables can be extracted from the association score, Patent Document 2 does not disclose a method for associating values having notation fluctuation held in the table.

  An object of the present invention is to appropriately generate a dependency relationship between item values in different documents.

  An analysis apparatus, an analysis method, and an analysis program that are one aspect of the invention disclosed in the present application are an analysis apparatus including a processor that executes the program and a storage device that stores the program, an analysis method using the analysis apparatus, And an analysis program to be executed by a processor accessible to the storage device. The storage device includes a calculation condition for defining a calculation range of a type of an item appearing in a document, a similarity calculation range between the value of the item of the type and another value of the item of the same type, and for corresponding to the type The classification information in which the classification conditions related to the plurality of pieces of feature information included in the item values are associated is stored. The storage device also includes a plurality of first feature information related to the value of the first item described in the first document and a plurality of second features related to the value of the second item described in the second document. Information. The processor refers to the classification condition of the classification information, identifies the type of the first item based on the plurality of first feature information, and based on the plurality of second feature information, The value type of two items is specified, the specifying process specifying the calculation condition corresponding to the type of the first item, the type of the first item specified by the specifying process, and the specifying process When the type of the second item is the same type, a calculation process for calculating the degree of similarity between the value of the first item and the value of the second item according to the calculation condition is calculated by the calculation process. And generating processing for generating dependency relationship information indicating a dependency relationship between the value of the first item and the value of the second item, using the similarity.

  According to the exemplary embodiment of the present invention, it is possible to appropriately generate the dependency relationship between the item values in different documents. Problems, configurations, and effects other than those described above will become apparent from the description of the following embodiments.

FIG. 1 is an explanatory diagram showing an example of design item classification in a file specification. FIG. 2 is an explanatory diagram illustrating an example of design item classification in an editing specification. FIG. 3 is an explanatory diagram illustrating an example of determining a similarity calculation method. FIG. 4 is an explanatory diagram showing a system configuration example of the analysis system. FIG. 5 is a block diagram illustrating a hardware configuration example of the analysis apparatus. FIG. 6 is a block diagram illustrating a functional configuration example of the analysis apparatus. FIG. 7 is an explanatory diagram of a generation example of the design item value storage table. FIG. 8 is an explanatory diagram showing an example of the design item classification table. FIG. 9 is an explanatory diagram of an example of the dependency relationship information. FIG. 10 is an explanatory diagram illustrating an example of weighted bipartite graph maximum matching by the selection unit. FIG. 11 is an explanatory diagram illustrating an example of a setting screen. FIG. 12 is an explanatory diagram illustrating an example of a dependency display screen. FIG. 13 is a flowchart illustrating an exemplary analysis processing procedure performed by the analysis apparatus. FIG. 14 is a flowchart illustrating a detailed processing procedure example of the similarity calculation processing (step S1308) between design item values illustrated in FIG.

<Example of analysis>
An analysis example of the dependency relationship between data will be described with reference to FIGS. Specifically, the analysis apparatus first determines that the first item described in the first document and the second item described in the second document of a different type from the first document are the same. Dependency relationship information indicating a dependency relationship between the value of the first item and the value of the second item is generated based on the similarity between the value of the item and the value of the second item. At that time, the analysis apparatus resembles (1) what index is used to classify the first item and the second item, and (2) what index is used to determine how to calculate the similarity. Determine before calculating the degree. (1) will be described with reference to FIGS. 1 and 2, and (2) will be described with reference to FIG.

  In the above description, a document is a document including tabular data such as a spread seed, and a character string indicating an item and a character string indicating an item value are described in a cell group of the tabular data. Has been. In the following description, a design document is used as an example of a document, and a design item is used as an example of an item. For example, the design document that is an example of the first document is, for example, a file definition specification, and the design document that is an example of the second document is an editing specification that is a different type from the file definition specification.

  FIG. 1 is an explanatory diagram showing an example of design item classification in a file specification. In FIG. 1, a file specification group DX includes one or more file specifications. A file specification is a design document that defines a file. The analysis apparatus acquires a design item and a design item value from each file specification in the file specification group, and generates a schema (first design item value storage table 101). In the example of FIG. 1, it is assumed that design items such as a file name, item name, and description are included as design items of the file specification, and columns Cxa, Cxb, Cxc,... Are generated.

  Specifically, for example, for an ordinary deposit file that is one of the file specifications, “ordinary deposit file” is stored in the column Cxa as the value of the file name that is a design item. Also, “Cxb1”, “Cxb2”, “Cxb3”,... Are stored in the column Cxb as item name values that are design items. Specifically, for example, “withdrawal date” is stored as the design item value Cxb1. Also, “Cxc1”, “Cxc2”, “Cxc3”,... Are stored in the column Cxc as explanation values that are design items. Specifically, for example, “the withdrawal date represents the date of withdrawal” is stored as the design item value Cxc1.

  The analysis device analyzes the value of the first item (a certain design item) described in the first document (a certain file specification), and generates a plurality of first feature information as the first analysis result 102. The plurality of first feature information is, for example, a ratio for each of a plurality of design item classification determination conditions for each column. The design item classification determination condition is a determination condition for classifying each design item, and is classified into the character type, the number of characters, and the presence or absence of punctuation marks. The character types are classified into groups such as “numbers only”, “alphanumeric characters only”, “alphanumeric characters and Japanese mixed”, for example. The number of characters is classified into groups such as “unified number of characters” and “undefined number of characters”. The presence / absence of punctuation marks is classified into groups such as “with punctuation marks” and “without punctuation marks”.

  Here, when the column Cxb is composed of six design item values Cxb1 to Cxb6, the character type, the number of characters, and the ratio of the presence or absence of punctuation marks will be specifically described.

  For example, in the case of “numbers only”, if the setting item values consisting only of numbers are Cxb1 and Cxb2, the column Cxb ratio Rxb1 of “numbers only” is Rxb1 = 2/6. The ratios Rxb2 and Rxb3 are similarly calculated for “alphanumeric only” and “mixed alphanumeric and Japanese”.

  In the case of “unification of the number of characters”, the analyzer generates a histogram of the number of characters, and the occupation ratio of the maximum number of the same characters is the ratio. The numbers of characters of the design item values Cxb1 to Cxb6 are 5, 2, 5, 3, 5, and 5, respectively. Since there are four characters, the maximum number of identical characters is five. Accordingly, the column Cxb ratio Rxb4 of “unified number of characters” is Rxb4 = 4/6. Conversely, the column Cxb ratio Rxb5 of “the number of characters is indefinite” is Rxb5 = 2/6.

  In the case of “with punctuation marks”, if the setting item values including punctuation marks are Cxb1 and Cxb2, the column Cxb ratio Rxb6 of “with punctuation marks” is Rxb6 = 2/6. Conversely, the column Cxb ratio Rxb7 of “no punctuation” is Rxb7 = 4/6.

  The analysis apparatus refers to the design item classification table 103 and identifies the design item of the column as one of the classifications using a plurality of design item classification determination conditions for each column. The specified design item classification is referred to as a design item classification 802. In this example, the design item classification 802 is specified from, for example, an ID, a physical name (an alphanumeric string that can be recognized by a computer), a Japanese name, a numerical value, and a sentence. For example, the design item classification 802 of the column Cxa is “Japanese name”. The design item classification 802 is associated with a dependency establishment condition 804. The dependency relationship establishment condition 804 is a calculation condition that defines a calculation range of the degree of similarity between the value of an item belonging to the design item classification 802 and another value of the same type of item. The calculation range includes, for example, “complete match only”, “allow similarity”, and “not applicable”.

  “Only complete match” indicates that the dependency relationship is established only when both design item values completely match. “Allow similarity” indicates that the dependency relationship is established when both design item values are in a similar relationship. “Not applicable” indicates that the dependency relationship between the two design item values is not obtained. For example, the dependency establishment condition for the design item classification “Japanese name” in the column Cxa is “allow similarity”. Therefore, a similar range of “Japanese name” is searched.

  In this way, the design item classification is specified for the design items (file name, item name, description,...) In the file specification.

  FIG. 2 is an explanatory diagram illustrating an example of design item classification in an editing specification. In FIG. 2, the editing specification group DY includes one or more editing specifications. An edit specification is a design document for editing a file definition. As for the edit specifications, the analysis apparatus acquires design items and design item values from each file specification of the file specification group, and generates a schema (second design item value storage table 201). Further, the analysis apparatus analyzes the value of the second item (a certain design item) described in the second document (a certain editing specification), and generates a plurality of second feature information as the second analysis result 202. Further, the analysis device refers to the design item classification table 103 and identifies the design item classification using a plurality of design item classification determination conditions for each column. In this way, the design item classification is specified for the design items (output file name, output item, editing condition,...) In the editing specification as in FIG.

  FIG. 3 is an explanatory diagram illustrating an example of determining a similarity calculation method. In the determination example of the similarity calculation method, the analysis apparatus uses the design document-design process correspondence table 300 and the design document relationship classification table 310 to generate the design item value dependency storage table 320. The design document-design process correspondence table 300 is a table in which a design document and a design process are associated with each other. The design process is a field indicating in which process the design document is created. The design document is a field indicating what kind of document it is. With respect to the file specification group shown in FIG. 1, the design document corresponds to the “file specification document”, so the design process is “detailed design document”. For the edit specification group in FIG. 2, the design document corresponds to “edit specification”, so the design process is “detailed design document”.

  The design document relationship classification table 310 is a table in which the design document relationship classification 311 and the similarity calculation formula 312 are associated with each other. The design document relation classification 311 has a relation source 311a and a relation destination 311b. That is, the similarity calculation formula 312 is determined by the combination of the design document relation classification 311 of the relation source 311a and the relation destination 311b. Since both the file specification and the editing specification correspond to the “detailed design specification”, the combination of the related source (for example, the file specification) 311a and the related destination (for example, the editing specification) 311b is the similarity calculation formula 312. This corresponds to “character string similarity”.

  Then, the analysis apparatus specifies the design item of the relation source and the value thereof, the similarity calculation formula 312 applied to the design item of the relation destination and the value thereof, and generates the design item value dependency storage table 320. . The design item value dependency storage table 320 includes a related source design item 321, a related source design item value 322, a related destination design item 323, a related destination design item value 324, a similarity calculation formula 325, and a similarity. 326 are stored.

  For example, the related source design item 321 of the entry on the fourth line is a file name (Cxa), and the related source design item value 322 is “foreign currency time deposit file”. The related design item 323 is an output file name (Cya), and the related design item value 324 is “foreign currency time deposit file”. Since both design items Cxa and Cya are included in the file specification and the editing specification, respectively, the corresponding similarity calculation formula 325 is “character string similarity”. Further, the design item classification of both the design items Cxa and Cya is “Japanese name” as described in FIGS. 1 and 2, and the dependency establishment condition is “allow similarity”. Therefore, the analysis apparatus associates the “ordinary deposit file” that is the related source design item value 322 of the file name (Cxa) that is the related source design item 321 and the output destination (Cya) that is the related destination design item 323. The similarity 326 with the “ordinary deposit file” that is the design item value 324 is calculated using the “character string similarity” that is the similarity calculation formula 325.

  In addition, when the dependency relationship establishment condition is “complete match only”, the maximum similarity is stored in the similarity 326 only when the related source design item value 322 and the related destination design item value 324 completely match. When the dependency relationship establishment condition is “not applicable”, the similarity calculation of the related source design item value 322 and the related destination design item value 324 is not executed, and the similarity 326 is blank.

  If the design item classification matches, an entry in the inter-design item value dependency storage table 320 is similarly generated for other design items. In this way, the analysis apparatus associates design item values among a plurality of sets in a set of design item values in which there is notation fluctuation. Therefore, items having a dependency relationship between documents can be associated with high accuracy.

  Although FIG. 3 illustrates the design process, the document classification is not limited to this. For example, when two vendors with different terms to use create and share a design document, the creation vendor may be used as an index for document classification. Furthermore, when organizations with different corporate cultures perform management integration and mergers, there may be cases where there are information system design documents held by each organization. In this case, an organization that owns the design document may be classified as a document. The design process, creation vendor, owned organization, and other document classification indexes listed above may be used alone, or may be defined as a composite condition of a plurality of classifications.

<System configuration example>
FIG. 4 is an explanatory diagram showing a system configuration example of the analysis system. The analysis system 400 includes an analysis device 401, a database (DB) 402, and a management terminal 403. The analysis device 401, the DB 402, and the management terminal 403 are communicably connected via a network 404 such as the Internet, a LAN (Local Area Network), and a WAN (Wide Area Network). The DB 402 stores a document group to be analyzed (for example, a file specification group and an editing specification group). The management terminal 403 operates the analysis device 401 via the network 404.

<Hardware configuration example>
FIG. 5 is a block diagram illustrating a hardware configuration example of the analysis apparatus 401. The analysis apparatus 401 includes a processor 501, a storage device 502, an input device 503, an output device 504, and a communication interface (communication IF 505). The processor 501, the storage device 502, the input device 503, the output device 504, and the communication IF 505 are connected by a bus. The processor 501 controls the analysis device 401. The storage device 502 serves as a work area for the processor 501. The storage device 502 is a non-temporary or temporary recording medium that stores various programs and data. Examples of the storage device 502 include a ROM (Read Only Memory), a RAM (Random Access Memory), a HDD (Hard Disk Drive), an SSD (Solid State Drive), and a flash memory. The input device 503 inputs data. Examples of the input device 503 include a keyboard, a mouse, a touch panel, a numeric keypad, and a scanner. The output device 504 outputs data. Examples of the output device 504 include a display and a printer. A communication IF 505 is connected to the network 404 to transmit / receive data. Examples according to the present invention will be described below.

<Functional configuration example>
FIG. 6 is a block diagram illustrating a functional configuration example of the analysis apparatus 401. The analysis apparatus 401 includes a design item value storage table 601, an analysis result storage table 602, a design item classification table 103, a design document-design process correspondence table 300, a design document relationship classification table 310, and dependency between design item values. The relationship storage table 320 is stored. Specifically, these tables are stored, for example, in the storage device 502 shown in FIG. Further, these tables may be stored in the DB 402.

  The design item value storage table 601 is a table for storing the first design item value storage table 101 and the second design item value storage table 201 shown in FIGS. 1 and 2. The analysis result storage table 602 is a table that stores the first analysis result 102 and the second analysis result 202 shown in FIGS. 1 and 2. The design item classification table 103 is a table for classifying design items into one of the design item classifications (described later in FIG. 8).

  The analysis apparatus 401 includes an acquisition unit 611, an analysis unit 612, a specification unit 613, a determination unit 614, a specification unit 613, a determination unit 614, a calculation unit 615, a generation unit 616, and a selection unit 617. And an output unit 618. Specifically, these are functions realized by causing the processor 501 to execute a program stored in the storage device 502, for example.

  The acquisition unit 611 acquires a document from the DB 402. The acquisition unit 611 groups the acquired documents. Specifically, for example, the acquisition unit 611 classifies the acquired document into a file specification group DX and an edit specification group DY. Regarding the classification, the acquisition unit 611 may perform automatic classification with reference to the file name, or may be classified according to the operation of the input device 503 or the management terminal 403. The acquisition unit 611 extracts a column from the acquired document and generates a design item value storage table 601 for each classified group. As a result, the first design item value storage table 101 shown in FIG. 1 and the second design item value storage table 201 shown in FIG. 2 are generated.

  FIG. 7 is an explanatory diagram showing a generation example of the design item value storage table 601. (A) shows a generation example of the first design item value storage table 101 from a certain file specification Dx, and (B) shows a generation example of the second design item value storage table 201 from a certain editing specification Dy. Show. Thick arrows indicate data extraction.

  Returning to FIG. 6, as shown in FIGS. 1 and 2, the analysis unit 612 analyzes an item value (for example, a design item value) and analyzes a plurality of pieces of feature information (first analysis result 102, first item). 2 analysis result 202), and the analysis result is stored in the analysis result storage table 602. The analysis result indicates feature information of the design item. For example, as described above, the analysis result is a ratio for each of the plurality of design item classification determination conditions calculated for each column.

  Thus, by providing a document group to the analysis apparatus 401, the process from the first feature information and the second feature information to the generation of dependency relationship information is automatically executed. Therefore, the efficiency of dependency relationship analysis can be improved.

  The identifying unit 613 identifies the type of item value based on a plurality of pieces of feature information. Specifically, for example, the specifying unit 613 specifies the value type of the first item based on the plurality of pieces of first feature information. The specifying unit 613 specifies the value type of the second item based on the plurality of second feature information. Here, the plurality of feature information is a ratio for each of a plurality of design item classification determination conditions calculated for each column in the analysis unit 612. The plurality of first feature information is a plurality of ratios for each column shown in FIG. 1, and the plurality of second feature information is a plurality of ratios for each column shown in FIG. The item is, for example, a design item, and the item value is, for example, a design item value. The item value type is a design item classification. In the example of FIG. 1, the specifying unit 613 specifies “Japanese name” from a plurality of design item classifications for the design item of the column Cxa (file name).

  The specifying unit 613 specifies a calculation condition corresponding to the type of the first item. Specifically, for example, the specifying unit 613 specifies the dependency establishment condition 804 that is a calculation condition corresponding to the design item classification in which the design items for the file specification are classified. In the example of FIG. 1, for the design item of the column Cxa (file name), the specifying unit 613 sets “acceptance of similarity” associated with “Japanese name” that is the design item classification 802 as the dependency establishment condition 804. As specified.

  The identifying unit 613 refers to the design item classification table 103 and identifies the type of item value and the calculation condition. Here, the design item classification table 103 will be described.

  FIG. 8 is an explanatory diagram showing an example of the design item classification table 103. The design item classification table 103 is classification information that identifies which design item classification a target design item corresponds to. The design item classification table 103 has, as fields, a design item classification ID 801, a design item classification 802, a design item classification determination condition 803, and a dependency establishment condition 804. The design item classification ID 801 is identification information that uniquely identifies the design item classification 802. The design item classification 802 indicates the type of design item that appears in the document.

  The design item classification determination condition 803 indicates a classification condition related to a plurality of pieces of feature information (each ratio of analysis results) possessed by the value of the design item for corresponding to the design item classification 802. Specifically, for example, as illustrated in FIGS. 1 and 2, the design item classification determination condition 803 is configured by a combination of feature information for a plurality of types of features. Here, as an example, it is composed of feature information about three types of features: a character type 831, the number of characters 832, and the presence or absence of punctuation marks 833. The dependency relationship establishment condition 804 indicates a calculation condition that defines the calculation range of the similarity 326 between the value of the design item of the design item classification 802 and another value of the same type of design item.

  Here, an example in which the design item classification 802 is specified using the design item classification table 103 will be described with reference to FIG. For example, the ratio of the column Cxa is seven types of Rxa1 to Rxa7. Among these, the ratio regarding the character type 831 is three types of feature information Rxa1 to Rxa3. For example, the specifying unit 613 applies the character type having the maximum ratio among Rxa1 to Rxa3 to the design item classification determination condition 803 of the column Cxa. Further, for example, a threshold is set in the design item classification determination condition 803, and the specifying unit 613 selects a character type having a ratio exceeding the threshold among Rxa1 to Rxa3 as the design item classification determination condition 803 for the column Cxa. You may apply to. Further, the specifying unit 613 may apply a character type that satisfies both to the design item classification determination condition 803 of the column Cxa.

  The ratio relating to the number of characters 832 (unified or indefinite) is two types of feature information Rxa4 and Rxa5. For example, the specifying unit 613 applies the number of characters having the maximum ratio among Rxa4 and Rxa5 to the design item classification determination condition 803 of the column Cxa. Further, for example, a threshold value is set for each design item classification determination condition, and the specifying unit 613 determines the number of characters that exceeds the threshold value among Rxa4 and Rxa5 as the design item classification determination condition 803 for the column Cxa. You may apply to. In addition, the specifying unit 613 may apply the number of characters satisfying both to the design item classification determination condition 803 of the column Cxa.

  The ratio regarding the presence or absence of punctuation marks 833 is two types of feature information Rxa7 and Rxa8. For example, the specifying unit 613 applies the number of characters having the maximum ratio among Rxa7 and Rxa8 to the design item classification determination condition 803 of the column Cxa. Further, for example, a threshold value is set in the design item classification determination condition 803, and the specifying unit 613 determines the number of characters having a ratio exceeding the threshold value among Rxa7 and Rxa8 as the design item classification determination condition 803 for the column Cxa. You may apply to. In addition, the specifying unit 613 may apply the number of characters satisfying both to the design item classification determination condition 803 of the column Cxa. Thus, the application range of similarity calculation is determined according to the design item classification 802.

  Returning to FIG. 7, the determination unit 614 calculates the similarity between the value of the first item and the value of the second item based on the combination of the first document type and the second document type. To decide. Specifically, for example, as illustrated in FIG. 3, the determination unit 614 determines the similarity calculation formula 325 using the design document-design process correspondence table 300 and the design document relationship classification table 310, and performs design. The item value dependency storage table 320 is generated.

  Here, the similarity calculation formula 325 includes, for example, “semantic similarity” and “character string similarity”. The “semantic similarity” is, for example, a similarity calculation method for calculating the similarity based on the direction and size of a co-occurrence vector that expresses the frequency of simultaneously used (co-occurring) words as a vector. If the difference in the direction and size of the co-occurrence vector between two design item values is within an allowable range, the two design item values are similar. “Character string similarity” is a similarity calculation method in which, for example, the similarity is calculated based on the edit distance that is the number of delete / add / replace operations required to set the difference between two character strings to 0. . If the difference in edit distance between two design item values is within an allowable range, the two design item values are similar.

  Here, an example of calculating the edit distance by the calculation unit 615 will be described. The similarity is a value obtained by subtracting from 1 a value obtained by dividing the edit distance between the two design item values by the average number of characters of the two design item values. For example, if both design item values are “Foreign Currency Time Deposit File” and “Foreign Currency Time Deposit File”, the “Foreign Currency Time Deposit File” can be changed to “Foreign Currency Time Deposit” by performing the deletion operation of “Foreign Currency” once. Matches "file". Therefore, the edit distance is “1”. Further, the average value of the number of characters (11) of the “foreign currency time deposit file” and the number of characters (10) of the “foreign currency time deposit file” is 10.5. A value obtained by dividing the edit distance between the two design item values by the average value of the number of characters of the two design item values is (1 ÷ 10.5). Therefore, the similarity is 0.90 (= 1− (1 ÷ 10.5)).

  When the type of the first item specified by the specifying unit 613 and the type of the second item specified by the specifying unit 613 are the same type, the calculation unit 615 calculates the first item value and the first item according to the calculation conditions. The similarity with the values of the two items is calculated. Specifically, for example, specifically, “Japanese name” that is the design item classification 802 of the design item (eg, file name) specified from the file specification and the specification specified from the editing specification, for example. The “Japanese name” that is the design item classification 802 of the design item (eg, output file name) is the same type. The dependency establishment condition 804 associated with “Japanese name” which is the design item classification 802 is “allow similarity”. In this case, the calculation unit 615 calculates the similarity 326 between the value “foreign currency time deposit file” of the design item (eg, file name) and the value “foreign currency time deposit file” of the design item (eg, output file name). calculate. That is, it is not required that both design item values completely match.

  When the determination unit 614 determines the similarity calculation method, the calculation unit 615 uses the determined similarity calculation method to calculate the value of the first item and the value of the second item according to the calculation condition. Calculate similarity. Thereby, prior to the calculation of the similarity 326 between the values of both items that are the generation targets of the dependency relationship, it is determined what specific similarity calculation method is appropriate.

  The calculation unit 615 does not calculate the similarity when the calculation condition is a condition that excludes the similarity calculation method. As a result, an increase in load due to useless similarity calculation can be suppressed, and the analysis process can be speeded up.

  In addition, when the calculation condition is a condition that defines a calculation range in which the values of both items whose similarity is to be calculated completely match, the calculation unit 615 determines that the value of the first item and the value of the second item are If they completely match, the similarity is set to a value indicating that the value of the first item is the same as the value of the second item. In addition, the calculation unit 615 does not calculate the similarity unless the value of the first item and the value of the second item completely match.

  Specifically, for example, it is assumed that the design item classification 802 is “ID” and the dependency establishment condition 804 is “complete match only”. The value “0001” of the design item (example: ID) completely matches the value “0001” of the design item (example: output ID). Instead of calculating using the similarity calculation method, the calculation unit 615 sets the similarity to a value (for example, the maximum value of similarity) indicating that both are the same. As a result, the similarity calculation load can be reduced, and the analysis process can be speeded up. On the other hand, the value “0001” of the design item (example: ID) does not completely match the value “0002” of the design item (example: output ID). In this case, since the calculation condition is “complete match only”, the calculation unit 615 does not calculate the similarity. As a result, an increase in load due to useless similarity calculation can be suppressed, and the analysis process can be speeded up.

  In addition, when the calculation condition is a condition that defines a calculation range that allows the values of both items of similarity calculation targets to be similar, the calculation unit 615 determines the value of the first item and the value of the second item. Is completely matched, the similarity is set to a value indicating that the value of the first item and the value of the second item are the same, and the value of the first item and the value of the second item are completely If they do not match, the similarity is calculated using the similarity calculation method.

  Specifically, for example, “Japanese name” which is the design item classification 802 of the design item (eg, file name) specified from the file specification, and the design item (eg, output file) specified from the edit specification “Japanese name” which is the design item classification 802 of “name” is the same kind. In this case, the calculation unit 615 completely matches the value “ordinary deposit file” of the design item (example: file name) and the value “ordinary deposit file” of the design item (example: output file name). Instead of calculating using the similarity calculation method, the calculation unit 615 sets the similarity to a value (for example, the maximum value of similarity) indicating that both are the same. As a result, the similarity calculation load can be reduced, and the analysis process can be speeded up.

  On the other hand, the design item (eg, file name) value “foreign currency time deposit file” and the design item (eg, output file name) value “foreign currency time deposit file” do not completely match. Thereby, the similarity calculation in the similar range is allowed. Therefore, it is possible to improve the accuracy of dependency analysis. Then, the calculation unit 615 calculates the similarity according to the similarity calculation method.

  The generation unit 616 generates dependency relationship information indicating the dependency relationship between the value of the first item and the value of the second item, using the similarity calculated by the calculation unit 615. Specifically, for example, the generation unit 616 uses the value of the first item and the value of the second item as nodes, connects the nodes with links, and assigns similarity to the links, thereby obtaining the dependency relationship information. Is generated. The generation unit 616 stores the generated dependency relationship information (related source design item 321 to similarity 326) in the design item value dependency storage table 320.

  When a certain design item value has a dependency relationship with a plurality of design item values in the dependency relationship information generated by the generation unit 616, the selection unit 617 determines the dependency relationship based on the similarity to the plurality of design item values. Select. Specifically, for example, the selection unit 617 selects a dependency using weighted bipartite graph maximum matching. As a result, the value of the second item having the optimum similarity to the value of the first item can be assigned to the value of the first item, and the dependency relationship analysis can be optimized. The weighted bipartite graph maximum matching will be described later.

  The output unit 618 outputs the dependency relationship information generated by the generation unit 616 and the selection result by the selection unit 617. Specifically, for example, the output unit 618 controls a display screen of a display that is an example of the output device 504, and displays dependency information and a selection result. In addition, the output unit 618 transmits dependency relationship information and selection results to the external device.

<Example of dependency information>
FIG. 9 is an explanatory diagram of an example of the dependency relationship information. The left side is a first node group 901 showing a related source design item value 322, and the right side is a second node group 902 showing a related destination design item value 324. A line segment connecting the nodes of the first node group 901 and the second node group 902 is a link connecting the nodes, and indicates that there is a dependency relationship between the design item values indicated by the nodes at both ends. The numerical value on the link is the similarity 326 between the two design item values connected by the link. In addition, the maximum value of the similarity 326 (that is, indicating perfect match) is set to “1”.

  Since the design item value “ordinary deposit file” of the node N9111 and the design item value “ordinary deposit file” of the node N921 completely match, the similarity 326 is “1”. Since the design item value “current account file” of the node N9112 and the design item value “current account file” of the node N922 completely match, the similarity 326 is “1”. Thereby, the dependency relationship is determined. The node group in the graph 903 surrounded by the dotted line is a node group that does not completely match, and the degree of similarity is less than 1. Therefore, the dependency relationship is not established.

  In the dependency relationship information, there is no link or similarity between design item values in the same document. That is, in the graph 903, the node group of the related source design item value 322 is the first subset 931, the node group of the related destination design item value 324 is the second subset 932, and any design item of the first subset 931 is displayed. It is a weighted bipartite graph in which a link and a similarity degree 326 exist between the value and any design item value of the second subset 932.

  Here, the selection unit 617 performs selection so that each node of the first subset 931 corresponds to only one node of the second subset 932 and the similarity 326 is maximized. Is called weighted bipartite graph maximum matching. The weighted bipartite graph maximum matching is the same as obtaining a correspondence relationship between the first subset 931 and the second subset 932 that maximizes the sum of the similarities 326. Thereby, the generation of the dependency relationship between the design item values can be automated.

  FIG. 10 is an explanatory diagram illustrating an example of a weighted bipartite graph maximum matching performed by the selection unit 617. The Hungarian method is known as a method for obtaining a weighted bipartite graph minimum weight maximum matching. In the following, the flow of solution by Hungarian law will be explained. The similarity matrix M1 is a matrix showing the similarity 326 of the combinations of design item values in the graph 903, with the related source design item values 322 as rows and the related destination design item values 324 as columns. In order to replace the problem of maximizing the similarity 326 with the minimization problem, each value of the similarity matrix M1 is multiplied by -1.

  In the Hungarian method, the selection unit 617 obtains the minimum value in each row for the similarity matrix M1, and subtracts the obtained value from all the values in the row. Further, the minimum value in each column is obtained, and the obtained value is subtracted from all the values in the column. A matrix obtained by converting the similarity matrix M1 by the above procedure is referred to as a post-conversion similarity matrix M2. In the similarity matrix M2 after conversion, the selection unit 617 determines whether or not an element having a value of 0 can be selected so as to be included in each row and each column. If such a selection can be made, the position of the element whose value is 0 in the converted similarity matrix M2 and the value of the similarity matrix M1 at the same position are the matching solution.

  The matrix M3 is a similarity matrix in which circles are added to values serving as matching solutions. As shown in the graph 1000, the node N913 (foreign currency time deposit file) and the node N923 (foreign currency time deposit file) are associated with each other by a link L33, and the matching obtained from the result is represented by the node N914 (property savings file). Node N925 (property deposit file) is associated with link L45, and node N915 (funding deposit file) and node N924 (funded fixed deposit file) are associated with link L54.

  Here, the node N913 (foreign currency time deposit file) and the node N923 (foreign currency time deposit file) are cases where fluctuations occur in “foreign currency” and “foreign currency denominated”. Node N 914 (property savings file) and node N 925 (property savings file) are cases where “saving” is omitted. The node N915 (funding deposit file) and the node N924 (funding fixed deposit file) are cases where “settlement” and “funding term” have a relationship between an abbreviated name and an official name. In both cases, there is a difference in notation, but because the same object is shown, it can be said that the objectives can be associated.

  In the above description, an example of modeling by bipartite graph minimum weight maximum matching has been shown, but for example, each design item value of the first subset 931 is between each design item value of the second subset 932. The degree of similarity 326 may be replaced with the degree of similarity ranking given by the size of the degree of similarity 326. In this way, a method may be used in which a similarity ranking is calculated and a matching is obtained so as to make the matching partner ranking as high as possible. For example, in the matching, which is obtained using the Gale-Chapley algorithm, even if the matching partner is recombined, it becomes a stable matching that cannot be recombined so that the similarity ranking is higher than the pasting partner. Is known (stable marriage problem), and the algorithm may be used.

<Example of display screen>
FIG. 11 is an explanatory diagram illustrating an example of a setting screen. The management terminal 403 acquires the design item classification table 103 stored in the storage device of the analysis apparatus 401 via the network 404. The management terminal 403 displays the acquired design item classification table 103 in the first area 1101 of the display setting screen 1100 of the management terminal 403. The second area 1102 displays an input field 1121 for the design item classification 802, a selection field 1122 to 1124 for the design item classification determination condition, and an input field for the dependency relationship establishment condition 804. Each input column and selection column 1121 to 1125 are input or selected via the input device of the management terminal 403.

  The management terminal 403 analyzes the information input or selected in each of the input fields and the selection fields 1121 to 1125 via the network 404 when the operation for inserting the registration button 1126 is accepted by the input device. To the device 401. As a result, the information is newly registered as a new record in the design item classification table 103 recorded in the storage device 502 of the analysis apparatus 401.

  In addition, the management terminal 403 receives the operation of selecting at least one of the design item classifications displayed in the first area on the input device and inserting the delete button 1127 as a trigger. The selection item and the deletion instruction are transmitted to the analysis apparatus 401 via the network 404. As a result, the record of the selected item is deleted from the design item classification table 103 recorded in the storage device 502 of the analysis apparatus 401.

  Accordingly, the design item classification determination condition 316 and the dependency relationship establishment condition 804 can be flexibly set according to the design document targeted by the analysis apparatus 401. In the above description, the editing operation of the design item classification table 103 has been described. However, the design document relationship classification 414 of the design document relationship classification table 310 can also be added, deleted, and modified. Further, addition, deletion, and correction may be made to the design item value storage table 601 in which values are automatically stored and the design item value dependency storage table 320.

  FIG. 12 is an explanatory diagram illustrating an example of a dependency display screen. Dependency relationship information, which is a correspondence between design item values obtained by a bipartite graph, is stored in a dependency relationship storage table 320 between design item values. Such dependency relationship information can also be used to visualize the dependency relationship between design item values. The management terminal 403 selects a file name 1112 as a starting point for tracing the dependency on the dependency relationship display screen 1200 of the display by operating the input device.

  The management terminal 403 retrieves the related design item value 324 in the inter-design item value dependency storage table 320 recorded in the storage device 502 of the analysis apparatus 401 when the operation for inserting the display button 1201 is accepted. Then, the related source design item value 322 in the same row is specified. Thereafter, based on the value of the specified related source design item value 322, the related destination design item value 324 of the inter-design item value dependency storage table 320 is searched to specify the related source design item value 322 in the same row. Repeat the process. As a result, the file 1211, the file 1212, the file 1213, and the file 1214 can be traced from the file 1210. By visualizing in this way, it becomes possible to promote understanding of the dependency relationship of design information.

  11 and 12, the example in which information is displayed on the display of the management terminal 403 by operating the management terminal 403 has been described. However, the display of the analysis apparatus 401 by operating the input device 503 of the analysis apparatus 401 is described. The information may be displayed on the screen.

<Analysis processing procedure example>
FIG. 13 is a flowchart illustrating an exemplary analysis processing procedure performed by the analysis apparatus 401. The analysis apparatus 401 uses the acquisition unit 611 to classify each design document acquired from the DB 402 for each document type, and generate a design item value storage table 601 as design information for each document type (step S1301).

  As shown in the first analysis result 102 in FIG. 1 and the second analysis result 202 in FIG. 2, the analysis apparatus 401 analyzes the column of design items in each design information by the analysis unit 612 and displays the design item. A plurality of ratios are calculated for each column (step S1302).

  The analysis apparatus 401 uses the specifying unit 613 to specify the design item classification 802 from the combination of ratios for each column of the design items as illustrated in FIGS. 1 and 2 (step S1303).

  The analysis apparatus 401 determines whether there is a combination of unselected identical design items by the determination unit 614 (step S1304). When there is an unselected combination (step S1304: Yes), the analysis apparatus 401 uses the determination unit 614 to select a combination of the same unselected related source design item 321 and related destination design item 323 (step S1305). .

  As shown in FIG. 3, the analysis apparatus 401 uses the determination unit 614 to identify the design process with reference to the design document-design process correspondence table 300 in the selected combination, and to include the design document including the related source design item 321. The design document relationship classification of the design document including the design document relationship classification and the related design item 323 is specified with reference to the design document relationship classification table 310 (step S1306).

  As illustrated in FIG. 3, the analysis apparatus 401 uses the determination unit 614 to calculate the similarity corresponding to the combination of the design document relationship classification 311 of the related source 311 a and the design document relationship classification 311 of the related destination 311 b specified in step S <b> 1306. The calculation formula 325 is determined with reference to the design document relation classification table 310 (step S1307).

  The analysis apparatus 401 executes a similarity calculation process between design item values using the identification result in step S1303 and the determination result in step S1307 (step S1308). Details of the similarity calculation processing between design item values (step S1308) will be described later with reference to FIG.

  The analysis apparatus 401 uses the generation unit 616 to generate a dependency relationship between design item values and store the dependency in the design item value dependency storage table 320 (step S1309). Then, the process returns to step S1304. If there is no unselected combination in step S1304 (step S1304: No), the analysis apparatus 401 uses the selection unit 617 to perform weighted bipartite graph maximum matching as shown in FIG. 10 (step S1310). . The analyzer 401 outputs the result of step S1310 (step S1310). Thereby, the analysis apparatus 401 ends a series of processes.

  FIG. 14 is a flowchart illustrating a detailed processing procedure example of the similarity calculation processing (step S1308) between design item values illustrated in FIG. The analysis apparatus 401 uses the calculation unit 615 to select an unselected related source design item value 322 from the selected related source design items 321 (step S1401). The analysis apparatus 401 uses the calculation unit 615 to select an unselected related destination design item value 324 from the selected related destination design items 323 (step S1402).

  The analysis apparatus 401 uses the calculation unit 615 to refer to the design item classification table 103 and depend on the design item classification 802 to which the selection related source design item value 322 (or selection related destination design item value 324) belongs. It is determined whether or not the fulfillment condition 804 is “not applicable” (step S1403).

  If it is “not applicable” (step S1403: Yes), the similarity 326 is not calculated, and the process proceeds to step S1408. On the other hand, if it is not “not applicable” (step S1403: No), the analysis apparatus 401 uses the calculation unit 615 to determine whether or not the selection related source design item value 322 and the selection related destination design item value 324 completely match. (Step S1404). When they completely match (step S1404: Yes), the similarity 326 between the selection related source design item value 322 and the selection related destination design item value 324 is set to the maximum value (step S1405), and the process proceeds to step S1408. That is, the similarity 326 is set without using the similarity calculation formula 325.

  On the other hand, if not completely matched (step S1404: No), the analysis apparatus 401 refers to the design item classification table 103 by the calculation unit 615 and selects the selection related source design item value 322 (or the selection related destination design item value 324). It is determined whether or not the dependency relationship establishment condition 804 associated with the design item classification 802 to which the item belongs is “allow similarity” (step S1406).

  If “similarity is allowed” (step S1406: Yes), the analysis apparatus 401 uses the similarity calculation formula 325 by the calculation unit 615 to select the selection related source design item value 322 and the selection related destination design item value 324. The similarity 326 is calculated (step S1407), and the process proceeds to step S1408.

  The analysis apparatus 401 determines whether there is an unselected related design item value 324 by the calculation unit 615 (step S1408). When there is an unselected related destination design item value 324 (step S1408: Yes), the process returns to step S1402. On the other hand, when there is no unselected related design item value 324 (step S1408: No), the analysis apparatus 401 determines whether there is an unselected related source design item value 322 by the calculation unit 615 (step S1408: No). S1409). When there is an unselected related source design item value 322 (step S1408: Yes), the process returns to step S1401. On the other hand, when there is no unselected related source design item value 322 (step S1408: No), the process proceeds to step S1309.

  As described above, the analysis apparatus 401 calculates, in the storage device 502, the degree of similarity between the type of item appearing in the document (design item classification 802) and the value of the item of the type and another value of the same type of item. A calculation condition (dependency establishment condition 804) that defines a range, a classification condition (design item classification determination condition 803) regarding a plurality of feature information (each ratio of analysis results) possessed by an item value corresponding to a type, Is stored (design item classification table 103). Further, the analysis apparatus 401 relates to the value (for example, “foreign currency fixed deposit file”) of the first item (for example, file name) described in the first document (for example, file specification) in the storage device 502. A plurality of first feature information (each ratio of the first analysis result 102) and a value (for example, “foreign currency”) of a second item (for example, an output file name) described in a second document (for example, an editing specification). A plurality of second feature information (each ratio of the second analysis result 202) related to the “time deposit file”) is stored.

  Then, the processor 501 refers to the classification condition of the classification information, specifies the type of the first item based on the plurality of first feature information, and determines the type of the second item based on the plurality of second feature information. (Design item classification 802) is specified, and calculation conditions corresponding to the type of the first item are specified. The processor 501 calculates the similarity 326 between the value of the first item and the value of the second item according to the calculation condition when the type of the first item and the type of the second item are the same type. To do. The processor 501 uses the similarity 326 to generate dependency relationship information indicating a dependency relationship between the value of the first item and the value of the second item.

  In other words, prior to calculating the similarity of the values of both items for which the dependency relationship is to be generated, the analysis device 401 has the same type of items according to the characteristics of the two items, and how to determine the type. Specific calculation conditions are applied. Therefore, the analysis device 401 can calculate the degree of similarity under conditions suitable for the values of both items according to the feature of the item, and can appropriately generate the dependency relationship between the item values.

  Further, the processor 501 determines the value of the first item and the value of the second item based on the combination of the type of the first document (for example, detailed design document) and the type of the second document (for example, detailed design document). The similarity calculation method (the semantic similarity or the character string similarity which is the value of the similarity calculation formula 312) is determined. The processor 501 uses the similarity calculation method to calculate the similarity between the value of the first item and the value of the second item according to the calculation condition.

  That is, the analysis apparatus 401 determines what kind of similarity calculation method is specifically appropriate prior to the calculation of the similarity 326 of the values of both items that are the generation targets of the dependency relationship. Therefore, the analysis apparatus 401 can calculate the similarity 326 using conditions and methods suitable for the values of both items according to the characteristics of the items, and can more appropriately generate the dependency between the item values. .

  Further, the processor 501 does not calculate the similarity 326 when the calculation condition is a condition that excludes the similarity 326 from the calculation target.

  Thereby, an increase in load due to useless similarity calculation is suppressed. Therefore, the efficiency of dependency relationship analysis can be improved.

  In addition, when the calculation condition is a condition that defines a calculation range in which the values of both items to be calculated of the similarity 326 completely match, the value of the first item and the value of the second item are complete. If they match, the similarity 326 is set to a value (for example, the maximum value of the similarity 326) indicating that the value of the first item and the value of the second item are the same, and the value of the first item and the second value If the value of the item does not completely match, the similarity 326 is not calculated.

  Thereby, an increase in load due to useless similarity calculation is suppressed. Therefore, the efficiency of dependency relationship analysis can be improved.

  Further, when the calculation condition is a condition that defines a calculation range in which the values of both items to be calculated of the similarity 326 are allowed to be similar, the value of the first item and the value of the second item Are completely matched, the similarity 326 is set to a value indicating that the value of the first item and the value of the second item are the same (for example, the maximum value of the similarity 326), and the value of the first item If the value of the second item and the value of the second item do not completely match, the similarity is calculated.

  Thereby, the similarity calculation in the similar range is allowed. Therefore, it is possible to improve the accuracy of dependency analysis.

  In addition, in the dependency relationship information, when the value of the first item has a dependency relationship with the plurality of values of the second item, the processor 501 determines the similarity between the value of the first item and each value of the second item. The value to be made dependent on the value of the first item is selected from the plurality of values of the second item.

  Thereby, the value of the second item having the optimum similarity to the value of the first item can be assigned to the value of the first item. Therefore, the dependency analysis can be optimized.

  In addition, the processor 501 analyzes the value of the first item described in the first document to generate a plurality of first feature information (first analysis result 102), and the second item described in the second document. The value of the item is analyzed to generate a plurality of second feature information (second analysis result 202) and stored in the storage device 502.

  Thus, by providing a document group to the analysis apparatus 401, the process from the first feature information and the second feature information to the generation of dependency relationship information is automatically executed. Therefore, the efficiency of dependency relationship analysis can be improved.

  The present invention is not limited to the above-described embodiments, and includes various modifications and equivalent configurations within the scope of the appended claims. For example, the above-described embodiments have been described in detail for easy understanding of the present invention, and the present invention is not necessarily limited to those having all the configurations described. A part of the configuration of one embodiment may be replaced with the configuration of another embodiment. Moreover, you may add the structure of another Example to the structure of a certain Example. Moreover, you may add, delete, or replace another structure about a part of structure of each Example.

  In addition, each of the above-described configurations, functions, processing units, processing means, etc. may be realized in hardware by designing a part or all of them, for example, with an integrated circuit, and the processor realizes each function. It may be realized by software by interpreting and executing the program to be executed.

  Information such as programs, tables, and files for realizing each function is recorded on a memory, a hard disk, a storage device such as an SSD (Solid State Drive), or an IC (Integrated Circuit) card, an SD card, a DVD (Digital Versatile Disc). It can be stored on a medium.

  Further, the control lines and the information lines are those that are considered necessary for the explanation, and not all the control lines and the information lines that are necessary for the mounting are shown. In practice, it can be considered that almost all the components are connected to each other.

103 Design Item Classification Table 300 Design Document-Design Process Correspondence Table 310 Design Document Relationship Classification Table 316 Design Item Classification Determination Condition 320 Design Item Value Dependency Storage Table 325 Similarity Calculation Formula 400 Analysis System 401 Analyzer 403 Management Terminal 404 Network 601 Design item value storage table 602 Analysis result storage table 611 Acquisition unit 612 Analysis unit 613 Identification unit 614 Determination unit 615 Calculation unit 616 Generation unit 617 Selection unit 618 Output unit 802 Design item classification judgment condition 804 Dependence condition establishment condition

Claims (9)

An analysis apparatus comprising a processor that executes a program and a storage device that stores the program,
The storage device includes a calculation condition for defining a calculation range of a type of an item appearing in a document, a similarity calculation range between the value of the item of the type and another value of the item of the same type, and for corresponding to the type A plurality of pieces of first feature information relating to a value of the first item described in the first document; A plurality of pieces of second feature information related to the value of the second item described in the two documents;
The processor is
With reference to the classification condition of the classification information, the type of the first item is specified based on the plurality of first feature information, and the value of the second item is determined based on the plurality of second feature information. A specifying process for specifying the type of the first item and specifying the calculation condition corresponding to the type of the first item;
When the type of the first item specified by the specifying process and the type of the second item specified by the specifying process are the same type, the value of the first item and the first item are determined according to the calculation condition. A calculation process for calculating the similarity with the values of the two items;
Generation processing for generating dependency relationship information indicating a dependency relationship between the value of the first item and the value of the second item, using the similarity calculated by the calculation processing;
The analysis apparatus characterized by performing. The analyzer according to claim 1,
The processor is
A determination process for determining a similarity calculation method for calculating a similarity between the value of the first item and the value of the second item based on a combination of the type of the first document and the type of the second document; Run,
In the calculation process, the processor calculates a similarity between the value of the first item and the value of the second item according to the calculation condition using the similarity calculation method determined by the determination process. An analyzer characterized by that. The analyzer according to claim 1,
In the calculation process, the processor does not calculate the similarity when the calculation condition is a condition that excludes the similarity from the calculation target. The analyzer according to claim 1,
In the calculation process, when the calculation condition is a condition that defines a calculation range in which the values of both items of the similarity calculation target completely match, the value of the first item and the second value If the value of the item completely matches, the similarity is set to a value indicating that the value of the first item and the value of the second item are the same, and the value of the first item and the second value If the value of the item does not completely match, the similarity is not calculated. The analyzer according to claim 1,
In the calculation process, when the calculation condition is a condition that defines a calculation range in which the values of both items of the similarity calculation target are allowed to be similar to each other, the value of the first item If the value of the second item completely matches, the similarity is set to a value indicating that the value of the first item is the same as the value of the second item, and the value of the first item is If the value of the second item does not completely match, the similarity is calculated. The analyzer according to claim 1,
The processor is
In the dependency relationship information generated by the generation process, when the value of the first item has a dependency relationship with a plurality of values of the second item, the value of the first item and each value of the second item An analysis apparatus that performs a selection process of selecting a value that depends on the value of the first item from a plurality of values of the second item based on the similarity. The analyzer according to claim 1,
The processor is
Analyzing the value of the first item described in the first document to generate the plurality of first characteristic information, analyzing the value of the second item described in the second document, and An analysis apparatus that generates a plurality of second characteristic information and executes an analysis process of storing the second characteristic information in the storage device. An analysis method by an analysis apparatus having a processor that executes a program and a storage device that stores the program,
The storage device includes a calculation condition for defining a calculation range of a type of an item appearing in a document, a similarity calculation range between the value of the item of the type and another value of the item of the same type, and for corresponding to the type A plurality of pieces of first feature information relating to a value of the first item described in the first document; A plurality of pieces of second feature information related to the value of the second item described in the two documents;
The processor is
With reference to the classification condition of the classification information, the type of the first item is specified based on the plurality of first feature information, and the value of the second item is determined based on the plurality of second feature information. A specifying process for specifying the type of the first item and specifying the calculation condition corresponding to the type of the first item;
When the type of the first item specified by the specifying process and the type of the second item specified by the specifying process are the same type, the value of the first item and the first item are determined according to the calculation condition. A calculation process for calculating the similarity with the values of the two items;
Generation processing for generating dependency relationship information indicating a dependency relationship between the value of the first item and the value of the second item, using the similarity calculated by the calculation processing;
The analysis method characterized by performing. An analysis program to be executed by a processor accessible to a storage device,
The storage device includes a calculation condition for defining a calculation range of a type of an item appearing in a document, a similarity calculation range between the value of the item of the type and another value of the item of the same type, and for corresponding to the type A plurality of pieces of first feature information relating to a value of the first item described in the first document; A plurality of pieces of second feature information related to the value of the second item described in the two documents;
In the processor,
With reference to the classification condition of the classification information, the type of the first item is specified based on the plurality of first feature information, and the value of the second item is determined based on the plurality of second feature information. A specifying process for specifying the type of the first item and specifying the calculation condition corresponding to the type of the first item;
When the type of the first item specified by the specifying process and the type of the second item specified by the specifying process are the same type, the value of the first item and the first item are determined according to the calculation condition. A calculation process for calculating the similarity with the values of the two items;
Generation processing for generating dependency relationship information indicating a dependency relationship between the value of the first item and the value of the second item, using the similarity calculated by the calculation processing;
An analysis program characterized in that

Images