Software requirement use case specification checking method, tool and computer program product

Abstract

This invention proposes a method, tool, and computer program product for checking the conformity of software requirement use cases. The checking method includes: extracting attributes from software requirement use cases; converting attributes into feature vectors using one-hot encoding based on a relevant vocabulary set and pre-defined weights for each attribute; ensuring the vector dimension equals the vocabulary size, with each vector position corresponding one-to-one with a word in the vocabulary set, and filling the corresponding vector position with a weighted average of word frequencies based on attribute weights; classifying software requirement use cases using a classification model based on the feature vectors; matching the corresponding writing standards according to the category of the software requirement use cases; and performing natural language structured analysis on the software requirement use case descriptions using an improved shift-reduction dependency analysis method to determine whether they conform to the writing standards. This invention solves the problem of low efficiency in manual checks, significantly improving check efficiency through multi-dimensional technical optimization, and promoting the standardization and intelligentization of software requirement review.

Description

Technical Field

[0001] This invention relates to the field of software requirements engineering technology, and in particular to a method, tool, and computer program product for checking the conformity of software requirements use cases. Background Technology

[0002] Software requirement use cases are core documents describing software functional requirements in the software development process. They bridge the gap between user needs and software implementation by clearly defining components such as use case names, executors, and scenario flows. However, in existing technologies, although the basic components of requirement use cases are clearly defined, the descriptions of these components often suffer from problems such as non-standardization, incompleteness, and ambiguity. For example, use case names may have disorganized structures, key information may be missing from success scenarios, and exception handling descriptions may be vague.

[0003] During the requirements review phase, relevant technical personnel need to spend a lot of time manually verifying the standardization of requirement use cases. This not only leads to low review efficiency and consumes a lot of manpower, but also makes it difficult to guarantee the comprehensiveness and accuracy of the inspection due to factors such as subjective human judgment and limited energy. This may lead to a series of problems in the subsequent software development process, such as misunderstanding of requirements, inconsistencies in functional implementation, and obstacles to testing and verification. Summary of the Invention

[0004] This invention provides a method, tool, and computer program product for checking the standardization of software requirement use cases, which solves the problems of low efficiency, incomplete coverage, and insufficient accuracy of manual checks in the prior art, and promotes the standardization and intelligentization of the software requirement review process.

[0005] The software requirement use case specification check method of this invention includes: Attributes are extracted from software requirement use cases, and one-hot encoding is used to transform the attributes into feature vectors based on the vocabulary set related to the requirement use cases and the pre-set weights of each attribute. The vector dimension is equal to the size of the vocabulary set, and each vector position corresponds one-to-one with a word in the vocabulary set. The word frequency is calculated according to the attribute weight and then filled into the corresponding vector position. Based on the feature vectors, the software requirement use cases are classified using a classification model; Based on the category of the software requirement use cases, the corresponding writing specifications are matched, and an improved shift-specification dependency analysis method is used to perform natural language structured analysis on the software requirement use case descriptions to determine whether they conform to the writing specifications.

[0006] The method of this invention can automate the entire process of requirement use case classification and standardization inspection, which can solve the problem of low efficiency of manual inspection and greatly improve the accuracy of classification and the reliability of standardization inspection.

[0007] According to some embodiments of the present invention, the method further includes: Identify the strongly related keywords for each software requirement use case category; If a strongly related word appears in the name of a software requirement use case, and the highest probability score of each category output by the classification model is lower than the confidence threshold or the difference between the highest probability score and the second highest probability score is less than the discrimination threshold, the category of the software requirement use case is determined to be the software requirement use case category corresponding to the strongly related word.

[0008] The strongly related words are operation verbs that are strongly associated with the category of the requirement use case, including add, add new, delete, query, modify, export, and count.

[0009] Therefore, by objectively determining the model's hesitation state through quantitative thresholds and combining category-specific strongly related words to assist in the determination, the problem of fuzzy model prediction can be accurately solved, further improving classification accuracy and avoiding the abuse or omission of auxiliary mechanisms.

[0010] According to some embodiments of the present invention, the classification model is selected based on the maximum entropy model.

[0011] Experiments have shown that this model outperforms other classification models such as support vector machines, logistic regression, and gradient boosting trees in demand use case classification scenarios.

[0012] The embodiments of this invention select the optimal model that is suitable for the classification scenarios of the required use cases, ensuring classification accuracy from the basic model level, avoiding classification deviations caused by improper model selection, and improving the overall classification effect.

[0013] According to some embodiments of the present invention, the extraction of attributes from software requirement use cases includes extracting use case name, context target, success scenario, exception handling, and alternative scenarios.

[0014] According to some embodiments of the present invention, the writing specifications are based on the requirements of coverage, sufficiency, correctness, consistency, designability, verifiability, clarity, and identifiability, and formulate writing guidelines that include multiple elements such as use case name, use case identifier, success scenario, alternative scenario, and exception handling.

[0015] Furthermore, the writing guidelines include: use case names should be short verb-object structures; use case identifiers should consist of English letters and hyphens and be unique; successful scenarios and alternative scenarios should include standardized operating steps without ambiguous words, and the steps should include actions, inputs, system behaviors, and processing results; exception handling should clearly define the corresponding steps, failure reasons, and system processing methods.

[0016] This provides a unified and practical standardized framework, eliminates ambiguity in rule writing, ensures the consistency and comprehensiveness of standard checks, and provides a clear basis for judging the quality of requirement use cases.

[0017] According to some embodiments of the present invention, the step of employing an improved shift-reduction dependency analysis method to perform natural language structured analysis on the software requirement use case description includes: The dependency relationship between verbs and objects is established by using commonly used verbs in the requirement use cases as a preset predicate candidate set and data elements in the use cases as a preset object candidate set.

[0018] The embodiments of this invention are specifically adapted to the language structure rules of "verb + data" in the required use cases, reducing syntactic analysis bias, improving the accuracy of extracting key information from use cases, and providing accurate structured data support for standardization checks.

[0019] This invention also proposes a software requirement use case specification check tool, comprising: The intelligent classification module is used to: extract attributes from software requirement use cases, and convert the attributes into feature vectors using one-hot encoding based on the vocabulary set related to the requirement use cases and the preset weights of each attribute; the vector dimension is equal to the size of the vocabulary set, each vector position corresponds one-to-one with a word in the vocabulary set, and the word frequency is calculated according to the attribute weight and then filled into the corresponding vector position; based on the feature vectors, the software requirement use cases are classified using a classification model. The test result filtering module is used to: match the corresponding writing specifications according to the category of the software requirement test cases; A specification detection module is developed to: use an improved shift-reduction dependency analysis method to perform natural language structured analysis on the software requirement use case descriptions and determine whether they conform to the writing specifications.

[0020] According to some embodiments of the present invention, a detection result filtering module is used to filter and visualize multiple detection results; The tool also includes: The detection result statistics module is used to perform statistical analysis on the detection results of requirement use cases from multiple dimensions and generate statistical reports.

[0021] This invention also proposes a computer program product, including a computer program that, when executed by a processor, implements the steps of the software requirement use case specification check method as described above.

[0022] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0023] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of the embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. In the drawings: Figure 1 This is a flowchart of the classification model training process in an embodiment of the present invention; Figure 2 This is a functional structure diagram of the software requirement use case specification check tool in this embodiment of the invention; Figure 3 This is a flowchart illustrating the relationship and processing flow of the various functional modules of the inspection tool in this embodiment of the invention; Figure 4 This is a schematic diagram of the main interface of the software requirements use case specification check tool according to an embodiment of the present invention. Detailed Implementation

[0024] Exemplary embodiments of the invention will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the invention are shown in the drawings, it should be understood that the invention can be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the invention and to fully convey the scope of the invention to those skilled in the art. Furthermore, in some instances, well-known methods, structures, and techniques have not been shown in detail so as not to obscure the understanding of this specification.

[0025] This invention aims to provide a method and tool for checking the standardization of software requirement use cases. By integrating machine learning and natural language processing technologies, it achieves automated classification, standardized specification verification, visual filtering of detection results, and multi-dimensional statistics of requirement use cases, thereby significantly improving the efficiency and accuracy of requirement use case standardization checks and providing reliable technical support for software requirement review.

[0026] The implementation process of the software requirement use case specification check method in this embodiment of the invention is as follows: A classification model is established using machine learning methods to automatically classify requirement use cases. Based on the type of requirement use case, an appropriate writing specification is selected, and natural language processing technology is used to perform specification checks on the requirement use cases. This method relies on the following three core technologies: classification model training, writing specification design, and specification check.

[0027] The classification model training steps aim to build a high-precision requirement use case classification model, laying the foundation for subsequent matching and specification writing. (Refer to...) Figure 1 As shown, the specific steps include the following: (1) Selection of Attributes and Weight Configuration for Requirement Use Cases: From multiple potential information of requirement use cases, five types of attributes with core influence on classification results were selected to avoid redundant attributes that would increase the complexity of model training and decrease accuracy. The selected attributes and their corresponding weights are shown in Table 1. Table 1. Selected Attributes and Weight Settings Use Case Name 3 Context target 1 Successful scenarios 2 Exception handling 1 Alternative scenarios 1 (2) Feature vector generation uses one-hot encoding to convert the selected requirement use case attributes into machine-recognizable feature vectors. The specific process is as follows: Construct a vocabulary set related to the requirements and use cases: The vocabulary set consists of all relevant words extracted from the above 5 types of attribute text, including operation verbs (such as add, query), data / object nouns (such as user, data), scenario / rule words (such as steps, exceptions), etc., and there are no duplicate words in the vocabulary set; Determine the vector dimension: The dimension N of the feature vector is equal to the size of the vocabulary set, and each position in the vector corresponds one-to-one with a single word in the vocabulary set; Weighted word frequency statistics and vector filling: Count the frequency of each word in the corresponding requirement use case attribute, and amplify the word frequency count according to the preset weight of the attribute (e.g., if "add" appears in the use case name, the weight is 3, so the word frequency count is multiplied by 3); fill the corresponding word position in the vector with the weighted word frequency, and fill the corresponding position of words that do not appear with 0, and finally form an integer feature vector.

[0028] (3) Classification model training and fine-tuning; The first step is model selection. Possible classification models include Support Vector Machine, Logistic Regression, Maximum Entropy Classification, and Gradient Boosting Tree. Experiments have shown that the Maximum Entropy model outperforms other models in the use case classification scenario; therefore, the Maximum Entropy model is selected as the core classification model. Then, identify the strongly related keywords. List the action verbs that are strongly associated with each use case category as strongly related keywords. For example, strongly related keywords for data addition use cases include: add, add new, increase, append, register, import, apply, receive, create, upload, build, establish, obtain, generate, etc. Finally, a model fine-tuning mechanism is constructed. The feature vector is input into the trained maximum entropy model, and the model outputs the probability score of the use case belonging to each category. If the prediction result has no discrimination, strongly related words are used to assist in the determination. If a strongly related word of a certain category appears in the name of the use case, the use case is directly determined to belong to that category. Here, "prediction has no discrimination" can be understood as the highest probability score of each category output by the model being lower than a preset confidence threshold, or the difference between the highest probability score and the second highest probability score being less than a preset discrimination threshold.

[0029] Software requirements documents are the requirements for the software product, transformed from user needs analysis by software engineers. They guide software design and verification. Therefore, software requirements must fulfill user requirements at the top, support software design at the bottom, and support software verification horizontally. IEEE 830-1998, "IEEE Recommended Practice for Software Requirements Specifications," emphasizes that requirements descriptions should avoid vague and ambiguous language and use accurate and verifiable terminology. Therefore, software requirements should meet the following requirements: 1. Coverage: Covers user requirements and is able to correctly cover user requirements.

[0030] 2. Sufficiency: Software requirements are not simply a repetition or decomposition of user requirements. They need to describe how the software product meets user requirements after thorough analysis, how the software makes the product easier for users to use, and whether it has the necessary fault tolerance for possible input errors and operations. These requirements must be reflected in the analysis.

[0031] 3. Correctness: The analysis results are correct.

[0032] 4. Coordination: The needs are coordinated and there are no contradictions among them.

[0033] 5. Designability: Supports software design, and the requirements for functional performance and general quality characteristics can be designed and implemented in the software.

[0034] 6. Verifiable: Supports verification of software functionality, performance, and general quality characteristics.

[0035] 7. Clarity: The description is clear and unambiguous.

[0036] 8. Identifiable: Each requirement should have a unique identifier to facilitate traceability.

[0037] The above requirements are translated into general writing guidelines that include multiple elements, as shown in Table 2: Table 2 General Guidelines for Writing Requirements Use Cases The requirement use case specification check is based on the trained classification model and preset writing specifications, and uses improved natural language processing technology to achieve automatic specification verification.

[0038] Software requirement use case specification checks involve syntactic analysis in natural language processing (NLP) technology. The previous section introduced NLP structure analysis methods, among which shift-reduction-based dependency parsing is simple to use and has low computational complexity, but the analysis results may be biased. This invention applies NLP structure analysis methods to requirement use case structure analysis and utilizes some heuristics to improve the method.

[0039] By analyzing the descriptions of the required use cases, we can conclude that the data elements in the use case descriptions are the objects of operations such as adding and modifying, and these objects function as objects in the sentence structure. Furthermore, the number of commonly used verbs corresponding to operations on data elements is very limited. Table 3 lists commonly used verbs in the requirement use case descriptions; these verbs typically function as predicates in the sentence structure. The accuracy of the analysis results can be improved through some auxiliary information. Therefore, this embodiment of the invention utilizes the information contained in the requirement use cases to optimize the shift-reduction dependency analysis method and applies it to the extraction of requirement use case information.

[0040] The core optimization of the improved shift-reduction dependency parsing method is as follows: by utilizing the linguistic rules of requirement use case descriptions (data elements are the operation objects, acting as objects; common verbs are the operation actions, acting as predicates), using the common verbs in requirement use cases listed in Table 3 as the preset predicate candidate set, and the data elements in the use cases as the preset object candidate set, the dependency relationship between predicates and objects is established first, thereby reducing syntactic analysis bias.

[0041] Table 3 Commonly used verbs in requirement use case descriptions 1 Modify, correct, select, change, edit, update, set, reset, approve, alter 2 Add, add, append, register, import, apply, receive, create, upload, build, establish, obtain, generate 3 Delete, Undo, Clear 4 Query, find, search, display, show, alarm, check 5 Export, download, forward, send, report, issue 6 Statistics, calculation, integration, measurement, summarization 7 Start, Pause, Restart, Terminate, Zoom In, Zoom Out, Move Up, Move Down, Move, Drag, Close The software requirement use case specification check method of the present invention has the following beneficial effects: High classification accuracy: By selecting core attributes and configuring reasonable weights, combined with the adaptability selection of the maximum entropy model and the fine-tuning mechanism assisted by strongly related words, the problem of model prediction ambiguity is effectively solved, the accuracy of requirement use case classification is greatly improved, and a reliable foundation is provided for subsequent specification matching.

[0042] Precise and accurate specification checks: Based on standardized writing guidelines and combined with an improved shift-specification dependency analysis method, it specifically adapts to the language structure rules of "verb + data" in requirement use cases, reduces natural language analysis bias, ensures the comprehensiveness and accuracy of specification checks, and avoids subjective omissions in manual checks.

[0043] High degree of automation: It automates the entire process of requirement use case classification, standardization testing, result filtering, and statistical report generation, significantly reducing the time cost of manual review and improving the efficiency of requirement review.

[0044] Highly user-friendly: The tool supports visual operation and targeted filtering, and the statistical reports are intuitive and easy to understand, reducing the user's operating threshold, adapting to the usage needs of users with different technical backgrounds, and promoting the implementation of standardized inspection processes.

[0045] Based on the above method, this invention has developed a supporting software requirement use case specification check tool, referring to... Figures 2-3 As shown, this tool is a visual operating software, and its core functions include four main modules: Intelligent classification module: Supports the training, testing and validation of classification models, and serializes and stores the trained models; receives user input of use cases, and realizes intelligent classification of use cases by calling the serialized classification model.

[0046] Specification Detection Module: Based on the use case categories output by the intelligent classification module, it automatically matches the corresponding writing specifications; it adopts an improved shift-specification dependency analysis method to parse and analyze the description of each element of the requirement use case, determine whether it conforms to the writing specifications, and generate detection results.

[0047] Results filtering module: Allows users to select specific rules according to their needs (such as filtering only "successful scenarios" and "anomaly handling" related rules) to perform targeted filtering of detection results, and displays the filtered results through a visual interface, making it easy for users to quickly view the target problem.

[0048] The statistical report module performs statistical analysis on the detection results from multiple dimensions such as violation type, use case category, and element dimension, and automatically generates statistical reports to intuitively present the overall standardization level of the requirement use cases.

[0049] The processing flow of this inspection tool includes: Users import requirement test cases to be inspected using the tool; The intelligent classification module automatically classifies requirement use cases; The standardization testing module matches the corresponding standards based on the classification results, performs standardization testing, and stores the test results in the backend document. Users can select target rules through the results filtering module and view the filtered results in a visual display. The statistics report module generates multi-dimensional statistics reports based on the full detection results, which users can download or view.

[0050] Figure 4 is a schematic diagram of the main interface of the software requirements use case specification check tool.

[0051] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific examples, so as to ensure that those skilled in the art can reproduce the present invention based on this embodiment.

[0052] I. Implementation Preparation Data preparation: Collect sufficient historical requirement use case data with labeled categories, covering multiple use case categories such as addition, deletion, query, and modification, as the training and test sets for the classification model; Vocabulary set construction: Extract relevant words from the core attribute text of the requirement use cases in the training set (use case name, context goal, success scenario, exception handling, alternative scenarios), and form a vocabulary set related to requirement use cases after deduplication; Threshold setting: Through experimental optimization on the test set, the preset confidence threshold for model prediction is 0.7 and the discrimination threshold is 0.1, which are used to determine whether the model is "hesitant". Tool Deployment: Deploy the developed software requirement use case specification check tool to the user terminal to ensure that the tool can call the classification model and write the specification library normally.

[0053] II. Implementation Steps for Classification Model Training Attribute extraction and weight assignment: For each requirement use case in the training set, extract five types of attributes: use case name, context goal, success scenario, exception handling, and alternative scenario, and assign weights of 3, 1, 2, 1, and 1 respectively; Feature vector generation: For the attribute text of each use case, count the frequency of each word in the attribute, amplify the values ​​according to their weights, and fill them into the vector positions corresponding to the vocabulary set to generate a feature vector; for example, if the vocabulary set is [add, query, user, data, step], and a use case name contains "add user" (weight 3), and a success scenario contains "step 1 input data" (weight 2), then the feature vector is [3, 0, 3, 2, 2]. Model training: Input the feature vectors and corresponding use case category labels into the maximum entropy model for model training, and adjust the model parameters until the classification accuracy on the test set meets the standard; Strongly related word configuration: Classify the commonly used verbs in Table 3 according to the use case category, for example, "add", "new", "register" are assigned to the "add" use case and stored as a strongly related word library; Model serialization: The trained model and a strongly related vocabulary are serialized and stored for use by tools.

[0054] III. Implementation Steps for Standardized Inspection Use case import: Users can import the requirement use case document to be inspected through the "File" button on the main interface of the tool (Figure 4); Intelligent classification: The tool calls the serialization model to extract attributes, generate feature vectors, and predict categories for imported test cases; if a test case model outputs "Add class 0.65, Query class 0.63" (highest score < 0.7, Top 2 difference < 0.1), and the test case name contains "Add", then it is determined to be an "Add class" test case; Specification matching: The tool automatically matches the writing specifications (based on Table 2 General Guidelines) corresponding to the "Add Class" use case. Structured analysis: An improved move-reduction dependency analysis method is used to identify predicates (such as "add") and objects (such as "user data") from use case descriptions, and to check whether the success scenario steps include actions, inputs, system behaviors and processing results, and whether there are ambiguous words, etc. Results output: The tool generates detection results, and users can filter and view the target results by checking options such as "use case name" and "success scenario" in the "rule selection" column; The statistical report module generates statistical information such as the distribution of violation types and the compliance rate of various elements, and users can export the reports.

[0055] Through the above implementation methods, the standardization of requirement use cases can be automated and accurately checked, effectively solving the problems of low efficiency, incomplete coverage and insufficient accuracy of manual inspection in the existing technology, and providing efficient and reliable technical support for software requirement review.

[0056] It should be noted that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention can have various modifications and variations. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0057] The contents not described in detail in this specification are common knowledge to those skilled in the art.

Claims

1. A software requirement use case specification checking method, characterized by, include: Attributes are extracted from software requirement use cases, and one-hot encoding is used to transform the attributes into feature vectors based on the vocabulary set related to the requirement use cases and the pre-set weights of each attribute. The vector dimension is equal to the size of the vocabulary set, and each vector position corresponds one-to-one with a word in the vocabulary set. The word frequency is calculated according to the attribute weight and then filled into the corresponding vector position. Based on the feature vectors, the software requirement use cases are classified using a classification model; Identify the strongly related keywords for each software requirement use case category; If a strongly related word appears in the name of the software requirement use case, and the highest probability score of each category output by the classification model is lower than the confidence threshold or the difference between the highest probability score and the second highest probability score is less than the distinction threshold, the category of the software requirement use case is determined to be the software requirement use case category corresponding to the strongly related word. Based on the category of the software requirement use cases, the corresponding writing specifications are matched, and an improved shift-specification dependency analysis method is used to perform natural language structured analysis on the description of the software requirement use cases to determine whether it conforms to the writing specifications. The improved shift-reduction dependency analysis method is used to perform natural language structured analysis on the software requirement use case description, including: The dependency relationship between verbs and objects is established by using commonly used verbs in the requirement use cases as a preset predicate candidate set and data elements in the use cases as a preset object candidate set.

2. The method of claim 1, wherein, The classification model selected is based on the maximum entropy model.

3. The method as described in claim 1, characterized in that, The extraction of attributes from software requirement use cases includes extracting use case name, context target, success scenario, exception handling, and alternative scenarios.

4. The method as described in claim 1, characterized in that, The writing guidelines are based on the requirements of coverage, sufficiency, correctness, consistency, designability, verifiability, clarity, and identifiability, and formulate writing guidelines that include multiple elements such as use case name, use case identifier, success scenario, alternative scenario, and exception handling.

5. The method as described in claim 4, characterized in that, The writing guidelines include: use case names should be short verb-object structures; use case identifiers should consist of English letters followed by hyphens and be unique; success scenarios and alternative scenarios should include standardized operating steps without ambiguous words, and the steps should include actions, inputs, system behaviors, and processing results; exception handling should clearly define the corresponding steps, reasons for failure, and system processing methods.

6. A software requirement use case specification check tool, characterized in that, include: The intelligent classification module is used to: extract attributes from software requirement use cases, and convert the attributes into feature vectors using one-hot encoding based on the vocabulary set related to the requirement use cases and the preset weights of each attribute; the vector dimension is equal to the size of the vocabulary set, each vector position corresponds one-to-one with a word in the vocabulary set, and the word frequency is calculated according to the attribute weight and then filled into the corresponding vector position; based on the feature vectors, the software requirement use cases are classified using a classification model; and the strongly related words of each software requirement use case category are identified. If a strongly related word appears in the name of the software requirement use case, and the highest probability score of each category output by the classification model is lower than the confidence threshold or the difference between the highest probability score and the second highest probability score is less than the distinction threshold, the category of the software requirement use case is determined to be the software requirement use case category corresponding to the strongly related word. The test result filtering module is used to: match the corresponding writing specifications according to the category of the software requirement test cases; A specification detection module is developed to: use an improved shift-reduction dependency analysis method to perform natural language structured analysis on the software requirement use case descriptions and determine whether they conform to the writing specifications; The improved shift-reduction dependency analysis method is used to perform natural language structured analysis on the software requirement use case description, including: The dependency relationship between verbs and objects is established by using commonly used verbs in the requirement use cases as a preset predicate candidate set and data elements in the use cases as a preset object candidate set.

7. The software requirement use case specification check tool as described in claim 6, characterized in that, The test result filtering module is used to filter and visualize multiple test results; The tool also includes: The detection result statistics module is used to perform statistical analysis on the detection results of requirement use cases from multiple dimensions and generate statistical reports.

8. A computer program product, characterized in that, It includes a computer program that, when executed by a processor, implements the steps of the software requirement use case specification check method as described in any one of claims 1 to 5.

Images