JP7555214B2 - Check item creation support device, check item creation support method, and program - Google Patents

Description

This disclosure relates to a check item generation support device, a check item generation support method, and a program that support the generation of check items for product specifications.

Technologies are known that identify inconsistencies in product specifications by linguistically analyzing the text of product specifications. For example, a specification verification device described in Patent Document 1 performs morphological analysis on required specifications written in natural language to break them down into words and parts of speech, and obtains clauses containing at least one word and syntax data representing dependency relationships between the clauses by dependency analysis. Next, the specification verification device analyzes the syntax data according to specification analysis rules to generate semantic expression data for each clause containing a verb word, the semantic expression data including items related to the verb, the subject of the verb, the object of the verb, and preconditions that must be satisfied before an action determined by the verb and the object is performed. The specification verification device then checks the semantic expression data to detect items that are incomplete in the semantic expression data.

JP 2014-81680 A

The specification verification device described in Patent Document 1 generates semantic expression data of the specifications by analyzing the syntax data obtained from the requirement specifications according to the specification analysis rules. Therefore, when the specification verification device described in Patent Document 1 is used to generate check items for product specifications, there is a problem that only check items with a predetermined sentence structure indicated by the specification analysis rules can be generated, and there is a possibility that items necessary for checking the product specifications may be missed.

The present disclosure aims to solve the above problem, and aims to provide a check item generation support device, a check item generation support method, and a program that can support the generation of check items for product specifications that are not limited to a predetermined sentence structure.

The check item generation support device disclosed herein includes a structure analysis unit that uses the results of analyzing the structure of a product specification to determine which parts of the specification are in table format or bullet point format ; a sentence generation unit that uses words, character strings or sentences included in the part of the specification determined by the structure analysis unit to generate sentences related to the product specifications corresponding to the part , and generates sentences for the specification to be processed including the generated sentences; a similarity analysis unit that obtains sentences corresponding to check items in stored data in which check items of the product specifications are accumulated, and sentences for the specification to be processed generated by the sentence generation unit , and calculates a similarity index that quantifies the degree of similarity between the two; and a reference information determination unit that extracts check items of the product specifications from the stored data based on the similarity index , and determines the extracted check items of the product specifications as reference information to be used in processing to generate check items of the product specifications from the specification to be processed.

According to the present disclosure, a similarity index is calculated that indicates the degree of similarity between a sentence corresponding to a check item in accumulated data in which check items of product specifications are accumulated and a sentence generated based on the structure of the specification to be processed, and check items extracted from the accumulated data based on the similarity index are determined as reference information to be used in the process of generating check items of the product specifications from the specification to be processed. The reference information is a check item extracted from the accumulated data without being limited to a predetermined sentence structure, and corresponds to a sentence similar to a sentence generated based on the structure of the specification to be processed. In this way, the check item generation support device according to the present disclosure can support the generation of check items of product specifications that are not limited to a predetermined sentence structure. Furthermore, by determining the reference information from the accumulated data, the quality of the check items can be stabilized, and it is possible to prevent the omission of items necessary for checking the product specifications.

1 is a block diagram showing a configuration of a check item generation support device according to a first embodiment; 4 is a flowchart showing a check item generation support method according to the first embodiment; FIG. 13 is a diagram showing an example of a table included in a product specification. FIG. 4 is a diagram showing an example of a sentence generated based on the table in FIG. 3 . FIG. 13 is a diagram showing an example of bullet points included in a product specification. FIG. 13 is a diagram illustrating an example of a target product description rule. FIG. 13 is a diagram showing an overview of a process for replacing words included in a sentence with a processing character string. FIG. 13 is a diagram showing an example of a sentence corresponding to a confirmation item in the stored data. FIG. 13 is a diagram showing an example of a sentence generated from a processing target specification. FIG. 1 shows an overview of similar analysis. FIG. 11A is a block diagram showing a hardware configuration for realizing the functions of the check item creation support device of embodiment 1, and FIG. 11B is a block diagram showing a hardware configuration for executing software for realizing the functions of the check item creation support device of embodiment 1.

Embodiment 1.
1 is a block diagram showing the configuration of a check item generation support device 1 according to a first embodiment. The check item generation support device 1 supports the process of generating check items of product specifications from a target specification by determining, from accumulated data 3 of check items of the product specifications, reference information used in the process of generating check items of product specifications from specification data 2 indicating the target specification. The specification data 2 is data indicating a target specification for which check items of the product specifications have not been generated. The accumulated data 3 is data in which check items that have a track record of ensuring product quality during product development are accumulated.

For example, the check item generation support device 1 extracts check items from the stored data 3 that correspond to sentences similar to sentences generated based on the structure of the specification to be processed, and determines the extracted check items as reference information. The check item generation support device 1 displays the reference information on the display device 4. Furthermore, if there are no check items in the stored data 3 that can be reference information, the check item generation support device 1 displays on the display device 4 that there is no reference information.

In FIG. 1, the check item generation support device 1 includes a structure analysis unit 11, a sentence generation unit 12, a similarity analysis unit 13, and a reference information determination unit 14. The structure analysis unit 11 defines the structure of a product specification by analyzing the structure of the specification. For example, the structure analysis unit 11 reads out specification data 2 indicating the specification to be processed, and analyzes the structure of the specification to be processed. The specification structure may be in a table format or a bullet-list format.

The sentence generation unit 12 generates sentences related to the product specifications based on the structure of the specifications to be processed. For example, when the structure of the specifications to be processed includes a table, the sentence generation unit 12 uses words included in the table to generate sentences related to the product specifications. In addition, when the structure of the specifications to be processed includes a list of multiple sentences and there is a relationship between the sentences in the list, the sentence generation unit 12 generates a sentence related to the product specifications by combining the individual related sentences.

In addition, the sentence generation unit 12 replaces words included in the sentence with processing strings according to target product description rules that associate words related to product specifications with processing strings. A processing string is a string that expresses the meaning of a word. For example, words such as "register", "memory", "signal", "bit" and "buffer" have the meaning of resources that define product specifications. The processing string corresponding to these is, for example, "$resource$". In other words, a word whose meaning is resource is replaced with the common processing string "$resource$" that expresses the same meaning.

In addition, the processing string corresponding to a word indicating a resource name, such as "A signal," is "$resource name$." In other words, a word whose meaning is a resource name is replaced with, for example, the common processing string "$resource name$" that expresses the same meaning.

The sentence generation unit 12 replaces words contained in sentences generated from the target specification document according to the target product description rules with processing character strings that correspond to their meanings, and replaces symbols contained in sentences that correspond to confirmation items in the stored data 3 with processing character strings that correspond to their meanings.

The similarity analysis unit 13 acquires a sentence corresponding to a check item in the stored data 3 and a sentence generated based on the structure of the specification to be processed, and calculates a similarity index that quantifies the degree of similarity between the two. For example, the similarity analysis unit 13 calculates the ratio of the number of words whose meanings are identical in both sentences corresponding to a check item in the stored data 3 (words replaced with processing strings) and sentences generated from the specification to be processed (words replaced with processing strings) to the total number of words in each sentence, and calculates the average ratio of each sentence as the similarity index between the two sentences.

The reference information determination unit 14 determines the check items of the product specifications extracted from the stored data 3 based on the similarity index as reference information to be used in the process of generating check items of the product specifications from the specification to be processed. For example, the reference information determination unit 14 reads out from the stored data 3 the check item that has the highest similarity index between the check items and the text generated from the specification to be processed, out of the check items stored in the stored data 3, and determines it as reference information.

FIG. 2 is a flowchart showing a check item generation support method according to the first embodiment.
First, the structure analysis unit 11 acquires specification data 2 indicating a specification to be processed, and analyzes the structure of the specification to be processed (step ST1) using the acquired specification data 2. The structure of the specification may be, for example, a table structure or a list structure.

Next, the sentence generation unit 12 generates sentences related to the product specifications based on the structure of the specifications to be processed (step ST2). For example, when a table is included in the structure of the specifications to be processed, the sentence generation unit 12 generates sentences related to the product specifications using the words included in this table. Figure 3 is a diagram showing an example of Table 2A included in a product specification, and shows Table 2A included in the specifications to be processed. The sentence generation unit 12 generates sentences related to the product specifications using the words included in Table 2A.

As shown in FIG. 3, in Table 2A, "register name," "address," "bit," and "function" are set as column information. If a value is described in the column corresponding to "function," the sentence generation unit 12 generates a sentence including the value description. The "value description" is a numerical value related to the "function" and its explanation. FIG. 4 is a diagram showing sentences A1 and A2 generated based on Table 2A in FIG. 3. Using the words described in the dashed lines in Table 2A, the sentence generation unit 12 generates sentence A1, "When the EN bit of the CONT register is 0, function 0 is disabled," and sentence A2, "When the EN bit of the CONT register is 1, function 0 is enabled."

When the structure of the specification to be processed contains bullet points and there is a relationship between the sentences in the bullet points, the sentence generation unit 12 generates a sentence that combines the individual related sentences as a sentence related to the product specifications. Figure 5 is a diagram showing an example of a bullet point included in a product specification. When the structure analysis unit 11 detects a sentence portion "The following conditions" in the specification to be processed, it predicts that sentences indicating the conditions will be listed following this sentence portion, and outputs an analysis result to the sentence generation unit 12 that the structure of the specification to be processed contains bullet points. Sentence 2B shown in Figure 5 consists of sentence 2b1, which contains the sentence portion "The following conditions", and sentences 2b2 and 2b3 that follow it.

When the sentence generation unit 12 detects the sentence portion "under the following conditions" from sentence 2b1 "A becomes B under the following conditions", it determines that sentences 2b2 and 2b3 following sentence 2b1 are sentences indicating conditions. Next, the sentence generation unit 12 determines that sentences 2b1 and 2b2 are itemized sentences and that there is a relationship between the sentences, and generates sentence B1, "A becomes B when C is 1", by combining sentences 2b1 and 2b2. Similarly, the sentence generation unit 12 determines that sentences 2b1 and 2b3 are itemized sentences and that there is a relationship between the sentences, and generates sentence B2, "A becomes B when D is in a valid state", by combining sentences 2b1 and 2b3.

Then, the sentence generation unit 12 replaces words included in sentences generated from the target specifications with processing strings corresponding to their meanings according to the target product description rules, and replaces symbols included in sentences corresponding to confirmation items in the stored data 3 with processing strings corresponding to their meanings. Figure 6 is a diagram showing an example of target product description rules 12A. The target product description rules 12A are rules that describe the correspondence between words related to the product specifications and processing strings that express the meanings of those words.

In FIG. 6, words such as "register", "memory", "signal", "bit" and "buffer" have the connotation of resources that define product specifications, so the processing strings corresponding to these words are "$resource$". Also, a word that has "alphanumeric characters" followed by the string "resource" corresponds to a processing string that represents a "resource name".

Figure 7 is a diagram showing an overview of the process of replacing words contained in sentence C1 with a character string for processing. The sentence generation unit 12 replaces the words contained in sentence C1, "The B register becomes 2 when the A signal is 1," with a character string for processing, in accordance with the target product description rule 12A. The words "register" and "signal" contained in sentence C1 are replaced with "$resource$" in accordance with the target product description rule 12A. As a result, sentence C2 is generated.

The sentence generation unit 12 replaces "B$resource$" with "$resourcename$" because "B" in "B$resource$" is an alphanumeric character indicating a resource name, and replaces "A$resource$" with "$resourcename$" because "A" in "A$resource$" is an alphanumeric character indicating a resource name. This generates sentence C3. Furthermore, "1" and "2" in sentence C3 have numerical meanings, so the processing character string corresponding to "1" and "2" is "$value$". The sentence generation unit 12 replaces "1" and "2" in sentence C3 with "$value$". This generates sentence C4.

In FIG. 2, the similarity analysis unit 13 acquires sentences corresponding to the check items in the stored data 3 and sentences generated based on the structure of the specifications to be processed, and calculates a similarity index between the two sentences (step ST3). FIG. 8 is a diagram showing examples of sentences corresponding to the check items in the stored data 3. FIG. 9 is a diagram showing examples of sentences generated from the specifications to be processed. The sentence generation unit 12 generates sentences C4a and C4b in which words are replaced with processing character strings, as shown in FIGS. 8 and 9, and outputs them to the similarity analysis unit 13.

As shown in FIG. 8, the similar analysis unit 13 divides sentence C4a into words and particles and generates sentence C5a from which the particles have been removed in order to analyze only the words contained in sentence C4a. Similarly, as shown in FIG. 9, the similar analysis unit 13 divides sentence C4b into words and particles and generates sentence C6a from which the particles have been removed in order to analyze only the words contained in sentence C4b.

Figure 10 is a diagram showing an overview of similar analysis. As shown in Figure 10, the similar analysis unit 13 calculates the number of matching words by comparing the words contained in sentence C5a and sentence C6a in order from the beginning of the sentence. Sentence C5a consists of "$resource name$", "$resource name$", "$value$", "when", "$value$", and "becomes", while sentence C6a consists of "$resource name$", "$value$", "when", "$resource name$", "$value$", and "becomes". The number of matching words between sentences C5a and C6a is "5".

Then, the similarity analysis unit 13 calculates the proportion Ia of words in sentence C5a that match the words in sentence C6a, calculates the proportion Ib of words in sentence C6a that match the words in sentence C5a, and calculates the average of these proportions as the similarity index I. In the example of FIG. 10, the number of words in sentence C5a is 6, and the number of words that match the words in sentence C6a is 5, so the proportion Ia is 5/6 = 0.8333. The number of words in sentence C6a is 6, and the number of words that match the words in sentence C5a is 5, so the proportion Ib is 5/6 = 0.8333. Therefore, the similarity index I is (0.8333 + 0.8333) / 2 = 0.8333.

The sentence generation unit 12 and the similarity analysis unit 13 execute the process from step ST2 to step ST3 for the sentences corresponding to all the check items in the stored data 3. This calculates a similarity index I between the sentences corresponding to each check item in the stored data 3 and the sentences generated from the target specification.

In FIG. 2, the reference information determination unit 14 determines the check items of the product specifications extracted from the stored data 3 based on the similarity index I as reference information to be used in the process of generating check items of the product specifications from the specification to be processed (step ST4). For example, the reference information determination unit 14 reads out from the stored data 3 the check item having the largest similarity index I between the check items stored in the stored data 3 and the text generated from the specification to be processed, and determines it as reference information. The reference information determination unit 14 may also extract check items having a similarity index equal to or greater than a certain threshold value from the stored data 3.

The reference information determination unit 14 extracts check items from the stored data 3 that correspond to sentences similar to sentences generated based on the structure of the specification to be processed, and determines the extracted check items as reference information. The reference information determination unit 14 displays the reference information on the display device 4. This allows check items that are similar to sentences generated from the specification to be processed and have a track record of ensuring product quality to be used in the process of generating check items for product specifications. This allows the check item generation support device 1 to support the generation of check items for product specifications that are not limited to a predetermined sentence structure. Furthermore, for example, if a similarity index I that exceeds a certain threshold is not calculated, the reference information determination unit 14 determines that there are no check items in the stored data 3 that can be reference information, and displays on the display device 4 that there is no reference information.

The hardware configuration for realizing the functions of the check item creation support device 1 is as follows.
Fig. 11A is a block diagram showing a hardware configuration for realizing the functions of the check item creation support device 1. Fig. 11B is a block diagram showing a hardware configuration for executing software for realizing the functions of the check item creation support device 1. In Fig. 11A and Fig. 11B, an input interface 100 is an interface that relays specification data 2 and accumulated data 3 input to the check item creation support device 1. An output interface 101 is an interface that relays reference information that is output to a display device 4.

The functions of the structure analysis unit 11, sentence generation unit 12, similarity analysis unit 13, and reference information determination unit 14 of the check item generation support device 1 are realized by a processing circuit. That is, the check item generation support device 1 has a processing circuit for executing the processes from step ST1 to step ST4 shown in FIG. 2. The processing circuit may be dedicated hardware, or may be a CPU (Central Processing Unit) that executes a program stored in memory.

When the processing circuit is the dedicated hardware processing circuit 102 shown in FIG. 11A, the processing circuit 102 is, for example, a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC (Application Specific Integrated Circuit), an FPGA (Field-Programmable Gate Array), or a combination of these. The functions of the structure analysis unit 11, the sentence generation unit 12, the similarity analysis unit 13, and the reference information determination unit 14 provided in the confirmation item generation support device 1 may be realized by separate processing circuits, or these functions may be realized together by a single processing circuit.

When the processing circuit is the processor 103 shown in FIG. 11B, the functions of the structure analysis unit 11, the sentence generation unit 12, the similarity analysis unit 13, and the reference information determination unit 14 provided in the confirmation item generation support device 1 are realized by software, firmware, or a combination of software and firmware. The software or firmware is written as a program and stored in the memory 104.

The processor 103 reads out and executes the programs stored in the memory 104 to realize the functions of the structure analysis unit 11, the sentence generation unit 12, the similarity analysis unit 13, and the reference information determination unit 14 of the check item generation support device 1. For example, the check item generation support device 1 includes a memory 104 that stores a program that, when executed by the processor 103, results in the processing of steps ST1 to ST4 shown in FIG. 2 being executed. These programs cause the computer to execute the procedures or methods of the structure analysis unit 11, the sentence generation unit 12, the similarity analysis unit 13, and the reference information determination unit 14. The memory 104 may be a computer-readable storage medium that stores programs for causing the computer to function as the structure analysis unit 11, the sentence generation unit 12, the similarity analysis unit 13, and the reference information determination unit 14.

The memory 104 may be, for example, a non-volatile or volatile semiconductor memory such as a RAM (Random Access Memory), a ROM (Read Only Memory), a flash memory, an EPROM (Erasable Programmable Read Only Memory), or an EEPROM (Electrically-EPROM), a magnetic disk, a flexible disk, an optical disk, a compact disk, a mini disk, or a DVD.

Some of the functions of the structure analysis unit 11, sentence generation unit 12, similarity analysis unit 13 and reference information determination unit 14 provided in the confirmation item generation support device 1 may be realized by dedicated hardware, and the remaining functions may be realized by software or firmware. For example, the structure analysis unit 11 has its functions realized by a processing circuit 102, which is dedicated hardware, and the sentence generation unit 12, similarity analysis unit 13 and reference information determination unit 14 have their functions realized by a processor 103 reading and executing a program stored in a memory 104. In this way, the processing circuit can realize the above functions by hardware, software, firmware or a combination of these.

As described above, the check item generation support device 1 according to the first embodiment calculates a similarity index indicating the degree of similarity between a sentence corresponding to a check item in the accumulated data 3 in which the check items of the product specifications are accumulated and a sentence generated based on the structure of the specification to be processed, and determines the check items extracted from the accumulated data 3 based on the similarity index as reference information to be used in the process of generating check items of the product specifications from the specification to be processed. The reference information is a check item extracted from the accumulated data 3 without being limited to a predetermined sentence structure, and corresponds to a sentence similar to a sentence generated based on the structure of the specification to be processed. In this way, the check item generation support device 1 can support the generation of check items of the product specifications that are not limited to a predetermined sentence structure. Furthermore, by determining the reference information from the accumulated data 3, the quality of the check items can be made uniform, and it is possible to prevent the omission of items necessary for checking the product specifications.

In addition, any of the components of the embodiments may be modified or omitted.

1 Check item generation support device, 2 Specification data, 2B, 2b1, 2b2, 2b3 Text, 3 Stored data, 4 Display device, 11 Structure analysis unit, 12 Text generation unit, 12A Target product description rules, 13 Similarity analysis unit, 14 Reference information determination unit, 100 Input interface, 101 Output interface, 102 Processing circuit, 103 Processor, 104 Memory.

Claims (6)

a structure analysis unit that uses a result of analyzing the structure of a product specification to identify a portion of the specification that is in a table format or a bulleted format ;
a sentence generation unit that uses a word, character string, or sentence included in the portion of the specification determined by the structure analysis unit to generate a sentence related to the product specification corresponding to the portion , and generates a sentence for the specification to be processed that includes the generated sentence ;
a similarity analysis unit that acquires a sentence corresponding to a check item in accumulated data in which check items of a product specification are accumulated, and a sentence about the specification to be processed generated by the sentence generation unit , and calculates a similarity index that quantifies the degree of similarity between the two;
a reference information determination unit that extracts product specification check items from the accumulated data based on the similarity index and determines the extracted product specification check items as reference information to be used in a process for generating product specification check items from the target specification;
A confirmation item generation support device comprising: the sentence generation unit replaces words included in the generated sentence with processing character strings in accordance with target product description rules in which words related to product specifications are associated with processing character strings that represent the meanings of the words;
The confirmation item generation support device according to claim 1, wherein the similarity analysis unit calculates a similarity index between a sentence in which words contained in a sentence corresponding to a confirmation item of the product specification extracted from the accumulated data are replaced with the processing character string, and a sentence in which words contained in a sentence generated for the specification to be processed are replaced with the processing character string . 3. The confirmation item generation support device according to claim 1, wherein, when the specification includes the text in tabular form , the text generation unit generates text related to the product specifications by using words included in the text in tabular form . The confirmation item generation support device according to claim 1 or 2, characterized in that, when the specification includes a plurality of sentences in the bullet -item format and there is a relationship between the sentences in the bullet-item format , the sentence generation unit generates a sentence that combines the related sentences as a sentence related to the product specifications. A structure analysis unit uses a result of analyzing the structure of a product specification to determine whether a part of the specification is in a table format or a bulleted format ;
a sentence generating unit using a word, character string or sentence contained in the portion of the specification determined by the structure analysis unit to generate a sentence related to the product specification corresponding to the portion , and generating a sentence for the specification to be processed including the generated sentence ;
A similarity analysis unit acquires a sentence corresponding to a check item in the accumulated data in which check items of the product specifications are accumulated, and a sentence about the specification to be processed generated by the sentence generation unit , and calculates a similarity index that quantifies the degree of similarity between the two sentences;
A reference information determination unit extracts product specification check items from the accumulated data based on the similarity index , and determines the extracted product specification check items as reference information to be used in a process for generating product specification check items from the target specification;
A confirmation item generation support method comprising: Computer,
a structure analysis unit that uses a result of analyzing the structure of a product specification to determine a portion of the specification that is in a table format or a bulleted format ;
a sentence generation unit that uses a word, character string, or sentence included in the portion of the specification determined by the structure analysis unit to generate a sentence related to the product specification corresponding to the portion , and generates a sentence for the specification to be processed that includes the generated sentence ;
a similarity analysis unit that acquires a sentence corresponding to a check item in accumulated data in which check items of a product specification are accumulated, and a sentence about the specification to be processed generated by the sentence generation unit , and calculates a similarity index that quantifies the degree of similarity between the two;
a reference information determination unit that extracts check items of product specifications from the accumulated data based on the similarity index , and determines the extracted check items of product specifications as reference information to be used in a process for generating check items of product specifications from the target specification;
A program to function as a

Images