A risk analysis method and system for software project management

Abstract

The present invention provides a risk analysis method and system for software project management, relating to the technical field of software project management. The method includes: collecting test record data of a target software project and a target test resource allocation strategy; extracting defect density features and test coverage features respectively corresponding to multiple functional modules; obtaining system architecture data of the target software project and constructing a module association graph of the target software project with respect to multiple functional modules; generating multiple resource allocation impact indexes for each functional module and processing them through a resource allocation association analysis model to generate a reference allocation strategy for the target software project; performing resource allocation risk analysis on the target test resource allocation strategy through the reference allocation strategy to generate a risk analysis result of the target software project with respect to the target test resource allocation strategy. The present invention realizes accurate analysis of the potential risks of the test resource allocation strategy.

Description

Technical Field

[0001] The present invention relates to the technical field of software project management, and in particular, to a risk analysis method and system for software project management. Background Art

[0002] Project management in the software development process is a key factor to ensure the success of a project. As the scale of software projects continues to expand and the development cycle becomes increasingly tight, the importance of the testing phase becomes more prominent. Software testing, as a key link affecting software quality, the limited nature of testing resources often becomes a bottleneck restricting project progress and quality assurance. If the testing resources are not properly allocated, there is a risk of project delay.

[0003] Some traditional testing resource allocation schemes rely on historical experience-driven. This experience-based allocation method has certain defects. In the case of limited testing resources, it is easy to lead to unbalanced resource allocation, thus increasing the risk of project failure. It is manifested that during the testing process, the testing of some function modules may be ignored due to improper resource allocation, and the defects of important modules may not be repaired in time due to insufficient resource allocation, etc. Conducting risk analysis on the allocation strategy of testing resources under limited testing resources can provide good suggestions for the reasonable allocation of testing resources, thereby improving testing efficiency. Summary of the Invention

[0004] To solve the above technical problems, the present invention proposes a risk analysis method and system for software project management, which can identify the risks of resource allocation in the testing phase of software projects and provide optimization suggestions for the reasonable allocation of testing resources.

[0005] To achieve the above object, in the first aspect of the present invention, a risk analysis method for software project management is provided, including:

[0006] Collecting test record data of a target software project regarding a target testing phase, and a target testing resource allocation strategy associated with the test record data;

[0007] Extracting defect density characteristics and test coverage characteristics respectively corresponding to multiple function modules in the target software project from the test record data;

[0008] Obtaining system architecture data of the target software project, and performing module association analysis on multiple function modules of the target software through the test record data and the system architecture data of the target software project to construct a module association map of the target software project regarding multiple function modules;

[0009] Generate multiple resource allocation impact indices for each functional module by means of a module association graph and the defect density characteristics and test coverage characteristics of multiple functional modules, and perform resource allocation processing on the resource allocation impact indices through a resource allocation association analysis model to generate a reference allocation strategy for the target software project, where the resource allocation association analysis model is constructed based on multiple sets of historical reference test record data;

[0010] Perform a resource allocation risk analysis on the target test resource allocation strategy through the reference allocation strategy to generate a risk analysis result of the target software project regarding the target test resource allocation strategy.

[0011] Preferably, extract the defect density characteristics and test coverage characteristics corresponding to multiple functional modules in the target software project from the test record data, including:

[0012] Determine the attribute data of each functional module, including the code size record data and the function point record data, extract the defect quantity characteristics of each functional module from the test record data, perform a defect distribution analysis on the code size record data through the defect quantity characteristics to generate the defect density characteristics of the functional module, extract the test coverage function point quantity characteristics of each functional module from the test record data, and perform a test coverage analysis on the function point record data through the test coverage function point quantity characteristics to generate the test coverage characteristics of the functional module.

[0013] Preferably, perform a module association analysis on multiple functional modules of the target software through the test record data and system architecture data of the target software project to construct a module association graph of the target software project regarding multiple functional modules, including:

[0014] Construct a module association graph of the target software project regarding multiple functional modules through the system architecture data, including using multiple functional modules as the nodes of the module association graph and determining the directed edges between the functional modules according to the system architecture data;

[0015] Separate the dependency record data between any two functional modules from the test record data, including data transmission record data and function call record data, extract the function call parameters and data transmission parameters of each directed edge in the module association graph from multiple sets of dependency record data, and calculate the module dependency parameters of each directed edge according to the function call parameters and data transmission parameters and add them to the module association graph.

[0016] Preferably, generate multiple resource allocation impact indices for each functional module by means of a module association graph and the defect density characteristics and test coverage characteristics of multiple functional modules, including:

[0017] Calculate the module association impact index of each functional module based on multiple module dependency parameters in the module association graph, respectively perform statistical analysis on the defect density characteristics and test coverage characteristics of multiple functional modules, generate the defect distribution impact index and test coverage impact index of each functional module, and extract the module scale impact index of the functional module from the attribute data of the functional module to obtain multiple resource allocation impact indexes of the functional module.

[0018] Preferably, perform resource allocation risk analysis on the target test resource allocation strategy through a reference allocation strategy, and generate a risk analysis result of the target software project regarding the target test resource allocation strategy, including:

[0019] Calculate the resource allocation deviation parameter of each functional module through the reference allocation strategy and the target test resource allocation strategy, perform test resource allocation risk analysis on the target software project based on multiple resource allocation deviation parameters, and calculate the resource allocation risk score of the target software project;

[0020] Among them, for the reference allocation strategy and the target test resource allocation strategy, they respectively include the resource allocation reference proportion parameter and the resource allocation target proportion parameter of each functional module.

[0021] Preferably, for the resource allocation association analysis model, it further includes:

[0022] Construct a training data set including multiple groups of sub-sample data through multiple groups of historical reference test record data. Each group of sub-sample data includes the historical resource allocation proportion parameter and multiple historical resource allocation impact indexes corresponding to each functional module in one group of the historical reference test record data;

[0023] Take the multiple historical resource allocation impact indexes in each group of sub-sample data as the training input of the resource allocation association analysis model, take the historical resource allocation proportion parameters corresponding to multiple groups of sub-sample data as the training target of the resource allocation association analysis model, and perform iterative training on the resource allocation association analysis model based on residual feature optimization through the sample data set, where the resource allocation association analysis model is a random forest model;

[0024] Performing iterative training on the resource allocation association analysis model based on residual feature optimization includes calculating the residual value of each group of sub-sample data after each round of iteration, performing convergence feature analysis according to the multiple residual values of each group of sub-sample data to calculate the convergence parameter of the sub-sample data after meeting the preset number of iterations, performing sample pruning processing on the sample data set based on the preset pruning parameter and the convergence parameter of the sub-sample data, and completing the training of the resource allocation association analysis model after reaching the preset convergence threshold.

[0025] The second aspect of the present invention provides a risk analysis system for software project management, which is used to execute the above-mentioned risk analysis method for software project management, including:

[0026] A software project data collection module, which is used to collect the test record data of the target software project regarding the target test phase, as well as the target test resource allocation strategy associated with the test record data;

[0027] A test feature extraction module, which is used to extract the defect density features and test coverage features corresponding to multiple functional modules in the target software project from the test record data;

[0028] A function association analysis module, which is used to obtain the system architecture data of the target software project, perform module association analysis on multiple functional modules of the target software through the test record data and system architecture data of the target software project, and construct a module association map of the target software project regarding multiple functional modules;

[0029] A test resource allocation analysis module, which is used to generate multiple resource allocation impact indexes for each functional module through the module association map and the defect density features and test coverage features of multiple functional modules, perform resource allocation processing on the resource allocation impact indexes through a resource allocation association analysis model, and generate a reference allocation strategy for the target software project, where the resource allocation association analysis model is constructed based on multiple groups of historical reference test record data;

[0030] A test risk analysis module, which is used to perform resource allocation risk analysis on the target test resource allocation strategy through the reference allocation strategy, and generate a risk analysis result of the target software project regarding the target test resource allocation strategy.

[0031] The present invention has the following beneficial effects:

[0032] By analyzing the test record data of the target software project, the present invention extracts features such as defect distribution and test coverage of different functional modules, constructs a module association map in combination with the system architecture data, analyzes the allocation impact relationship between functional modules and test resources from multiple dimensions, constructs a multi-dimensional resource allocation impact index, uses a model trained based on historical reference data to intelligently generate a resource allocation reference strategy, accurately identifies high-risk modules and evaluates the overall resource allocation risk by quantitatively comparing the resource allocation deviation between the target strategy and the reference strategy, realizes the dynamic matching of test resource allocation with module relevance, defect distribution and test coverage features, can provide an evaluation basis for the project team on whether the test resource allocation is reasonable, helps to discover potential resource allocation problems, and provides guidance for subsequent resource optimization. Description of the Drawings

[0033] Figure 1This is a schematic flowchart of a risk analysis method for software project management according to an exemplary embodiment of the present invention.

[0034] Figure 2 This is a schematic structural diagram of a risk analysis system for software project management according to an exemplary embodiment of the present invention. Detailed implementation manners

[0035] In order to enable those skilled in the art to better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0036] Please refer to Figure 1 , an embodiment of the present invention provides a risk analysis method for software project management, including the following steps:

[0037] Step S1: Collect the test record data of the target software project regarding the target test stage, as well as the target test resource allocation strategy associated with the test record data.

[0038] Specifically, for the test record data of the target software project in the target test stage, it can specifically be the relevant record data in a certain test stage where the software project is not comprehensively tested in the initial stage. In the initial stage of the overall test stage, generally, the basic functions of each module in the software system are initially verified to see if they work as expected, and whether the basic interactions between modules are normal, so as to verify the initial reliability after system integration. Relevant testers will formulate relevant target test resource allocation strategies based on the corresponding test record data to guide the subsequent process of optimizing or repairing different functional modules of the system. In this process, if the test resources are limited, including human resources such as testers and time resources such as test duration, if the allocation is unreasonable, it may lead to risks such as incomplete test coverage, insufficient testing of important modules, and lagging test progress, thereby affecting the quality and delivery cycle of the project. Conducting a comprehensive analysis of the collected test record data to perform a risk analysis on the target test resource allocation strategy can well discover potential risks in resource allocation in advance, facilitate optimizing resource utilization and test efficiency to improve the success rate of the entire project.

[0039] The test record data includes, but is not limited to, test-related information such as test cases used for different functional modules and corresponding test results. A functional module refers to a component in a software system that is responsible for a specific function or task. The test resource allocation strategy associated with these test records can be specifically formulated by the project team based on empirical knowledge. Subsequently, the test resource allocation-related data for further in-depth analysis and repair of different functional modules based on these results constitutes the data basis for the subsequent resource allocation risk analysis process.

[0040] Step S2: Extract the defect density characteristics and test coverage characteristics corresponding to multiple functional modules in the target software project from the test record data.

[0041] Specifically, for the collected test record data, extract the defect density characteristics regarding the defect distribution characteristics of different functional modules and the test coverage characteristics regarding the comprehensiveness of the test process from the test records. In this process, first determine the attribute data of each functional module, including information such as code size record data and function point record data, which are used to reflect the number of lines of code, the number of classes, etc. contained in different functional modules, and the number of function points that different functional modules need to implement, that is, the number of business functions completed by the module.

[0042] And further extract the defect quantity characteristics of each functional module from the test record data. Specifically, the defect report of the functional module regarding the test process can be extracted from the test record data to obtain the defect quantity characteristics corresponding to each module. Combine the defect quantity characteristics of the functional module and the code size record data, and conduct defect distribution analysis on the functional module. For example, take the ratio of the defect quantity characteristics and the code size characteristics corresponding to the functional module, such as the number of lines of code, as the defect density characteristic of the functional module, which characterizes the defect distribution characteristics of different functional modules. At the same time, also extract the test coverage function point quantity characteristics of each functional module from the test record data, such as the number of function points that have been tested, to evaluate the comprehensiveness of the test for different functional modules in the target test phase. Then, conduct test coverage analysis on the function point record data through the test coverage function point quantity characteristics, and take the ratio between the test coverage function point quantity characteristics and the total function points as the test coverage characteristic of the functional module, indicating the overall coverage degree of the test process for different functional modules.

[0043] Step S3: Obtain the system architecture data of the target software project, and conduct module association analysis on multiple functional modules of the target software through the test record data and system architecture data of the target software project to construct a module association graph of the target software project regarding multiple functional modules.

[0044] Specifically, in order to deeply mine the characteristics of the aforementioned test-related features, comprehensive analysis is carried out in combination with the functional association or dependency features among different functional modules in the target software project. In this process, first, based on the system architecture data of the target software project and the information recorded in the test record data during the test process, a module association graph of the target software project for multiple functional modules is constructed. System architecture data usually includes information such as interface dependencies, data flows, and call relationships between different functional modules. Through this information, it can help understand the nature of the associations and the scope of influence between different modules. And the test record data collected during the test process also contains the module dependency characteristics of some key functions during the test process. Constructing a module association graph based on these data can help identify which functional module defects may trigger a chain reaction and affect the normal testing or operation of other modules.

[0045] For the construction process of the module association graph, first, multiple functional modules are used as the nodes of the module association graph, and the directed edges between the functional modules are determined according to the system architecture data. Among them, the functional modules that are multiple nodes of the module association graph in the software system may involve independent modules, sub-modules, or service units in the software system. In the module association graph, the associations between functional modules are represented by directed edges, which reflect the dependency relationships or interaction effects between modules. The system architecture data can be used to determine the directed edges between each functional module. For example, if module A calls a certain function of module B, or module A transfers data to module B, these can all be represented by directed edges, and the direction of the arrow represents the direction of the dependency relationship, thus completing the preliminary construction of the module association graph.

[0046] And further improve the module association graph in combination with the test record data, and extract the actual quantitative dependency relationships. First, separate the dependency record data between any two functional modules from the test record data, which at least includes information such as data transmission record data and function call record data, describing the data flow relationship and function call relationship between functional modules during the test process. Quantitative indicators such as the call frequency and data transmission volume between functional modules can be extracted from it. The call frequency indicates, for example, the number of times module A calls the relevant function in module B during the test process, which can reflect the direct dependency relationship between the two modules. The higher the call frequency, the stronger the dependency between the modules. Specifically, the function call parameters between functional modules can be statistically generated by analyzing the function call record data. The data transmission volume indicates, for example, how much data module A transfers to module B, reflecting the data dependency relationship between the two modules. If a module sends data to another module with a high frequency or a large amount of data, it means that the dependency relationship between them is strong. By analyzing the data transmission record data of the module, the data transmission parameters regarding the data transmission characteristics between different functional modules can be statistically generated.

[0047] Finally, extract the function call parameters and data transfer parameters of each directed edge in the module association graph from multiple sets of dependency record data. Then, the module dependency parameters of each directed edge can be calculated based on the function call parameters and data transfer parameters. For example, for the directed edge between module A and module B, the module dependency parameter of the directed edge is obtained by performing weighted calculation on the corresponding function call parameters and data transfer parameters of the directed edge. Different weight parameters are used to adjust the importance of the call frequency and data transfer volume. If it is considered that the call frequency and data transfer volume have equal influence, the same weight parameter such as 0.5 can be adopted. Those skilled in the art can reasonably set it according to actual test needs. Finally, the calculated module dependency parameters are added to the module association graph. Each directed edge not only represents the relationship between modules, but also can reflect the importance of interaction between modules through the dependency strength. By combining the system architecture data and test record data, a complete module association graph is finally constructed, where each node represents a functional module, each directed edge represents the dependency relationship between modules, and this dependency strength is quantified by the module dependency parameter.

[0048] Step S4: Generate multiple resource allocation impact indices for each functional module through the module association graph, the defect density characteristics, and the test coverage characteristics of multiple functional modules, and perform resource allocation processing on the resource allocation impact indices through the resource allocation association analysis model to generate a reference allocation strategy for the target software project.

[0049] Specifically, determine multiple resource allocation impact indices of different functional modules regarding the strategies for influencing the defect repair and optimization of subsequent modules, or rather, measure the requirements of different modules for the overall test resources, based on the association and dependency characteristics between functional modules characterized by the module association graph and the problems existing in different functional modules during the preliminary test process.

[0050] In this process, for the module association graph, based on multiple module dependency parameters in the module association graph, module association impact analysis is performed on each functional module. Specifically, according to the degree characteristics involved in the functional modules in the module association graph, including in-degree and out-degree, which reflect the dependencies received by the functional module and the dependencies generated by itself, the in-degree and out-degree of the functional module can be summarized to obtain the overall degree characteristics between the functional module and the other modules. And based on the degree characteristics of multiple functional modules, the module association impact index of each functional module is calculated. For example, the proportion of the degree characteristics of the functional module in the whole is used as the module association impact index. The larger the module association impact index, that is, the more strongly associated modules, may require more resources to ensure that their defects are repaired as soon as possible or the modules are optimized, and the impact on the other modules is reduced as much as possible. Similarly, statistical analysis is respectively performed on the defect density characteristics and test coverage characteristics of multiple functional modules to generate the defect distribution impact index and test coverage impact index of each functional module. Functional modules with high defect density or low test coverage rate may require more resources for defect repair or supplementary test work, etc. By quantifying the proportion of the defect density characteristics and test coverage characteristics of the functional module in the whole, the defect distribution impact index and test coverage impact index of the functional module are calculated. At the same time, the scale characteristics of different functional modules also need to be considered. The module scale impact index of the functional module can be extracted from the attribute data of the functional module. The module scale reflects the complexity of the module and the relative size of the required test resources. The larger the module scale, the greater the impact on resource allocation. The module scale impact index is quantified by quantifying the ratio of the regular scale of different functional modules to the scale of the whole software system, which is used to represent the impact of the module scale on the test resource requirements. In this way, multiple resource allocation impact indexes of functional modules are generated from different levels to quantitatively evaluate the possible impact of functional modules on resource allocation.

[0051] After that, through the resource allocation association analysis model constructed based on multiple groups of historical reference test record data, these multi-dimensional impact indexes can be optimized, and the impact of these impact characteristics of the module on resource allocation can be analyzed to generate the reference allocation strategy for the target software project.

[0052] The resource allocation correlation analysis model is obtained by training with a training dataset constructed based on multiple sets of historical reference data. Specifically, a training dataset including multiple sets of subsample data is constructed through multiple sets of historical reference test record data. The historical reference test record data can be reference cases with good performance in terms of relevant resource allocation effects and the next-stage test in the same test stage as the target test stage in different historical software projects. Specifically, it can be evaluated by an expert team. For example, comprehensively analyze whether the defect repair progress or module optimization progress of each functional module is consistent, whether the key functional modules have achieved good repair effects, and whether there is a phenomenon of waste of test resources. Sample data with good performance in the target test stage in historical projects are selected as references through multi-angle analysis.

[0053] Each set of subsample data in the training dataset includes the historical resource allocation proportion parameters and multiple historical resource allocation impact indices corresponding to each functional module in one set of the historical reference test record data. Similar to the aforementioned multiple resource allocation impact indices, the historical resource allocation proportion parameters can specifically be the historical test resource allocation strategies for the historical reference test record data formulated by the project team in combination with the historical reference test record data and empirical knowledge, which are used to indicate the test resources allocated to different functional modules, including human resources, equipment resources, etc. For example, the number of members and equipment responsible for defect repair and further functional analysis of different functional modules. The proportion of different test resources in the overall resources is quantified through analysis.

[0054] After the training dataset is constructed, the multiple historical resource allocation impact indices in each set of subsample data are used as the training input of the resource allocation correlation analysis model. The model can learn the impacts of characteristics such as defect distribution, test coverage, and module association among functional modules in a software project on test resource allocation through these impact indices, and use the historical resource allocation proportion parameters corresponding to each set of subsample data as the training target of the resource allocation correlation analysis model, enabling the model to learn how to analyze the resource allocation proportion based on multiple impact indices, so as to optimize the resource allocation under the given impact indices.

[0055] In this embodiment, a random forest model is selected as the resource allocation correlation analysis model. To further improve the accuracy and convergence of the model, optimization based on residual features is used for iterative training of the model. During the training process, after each round of iteration, the residual values are calculated, and the prediction ability of the model is improved by analyzing these residual values.

[0056] Specifically, after each iteration, the residual values of each group of subsample data are calculated. The residual value refers to the difference between the actual value in the dataset and the predicted value output by the model, representing the error of the model in each round of training. Since it is considered that there are differences in modules in different historical software projects, and this solution uses the characteristics of different dimensions of functional modules as samples during the analysis process, there may be a situation where it is not easy to converge during the model training process. Therefore, residual analysis is combined to optimize the training process. Specifically, after meeting the preset number of iterations, convergence feature analysis is performed based on the multiple residual values of each group of subsample data to calculate the convergence parameters of the subsample data. In this process, the convergence trend of multiple residuals of the subsample data in different time windows can be analyzed through a sliding window. For example, if it conforms to the overall trend of gradually decreasing, it is recorded as 1, otherwise it is marked as 0, so as to extract multiple convergence trend values. Through the time distribution analysis of multiple convergence trend values, for example, the later the window, the higher the trend value score, an exponential decay function is constructed based on the change of the time window to determine the scoring weights of different convergence trend values, and the multiple convergence trend values are weighted to calculate the convergence parameters of the subsample data, which are used to evaluate the fit degree between the subsample data and the current model. The larger the convergence parameter, the higher the fit degree. And based on the preset pruning parameter and the convergence parameter of the subsample data, sample pruning is performed on the sample dataset. The preset pruning parameter can be based on the percentile of the number of subsample data, such as 90%, that is, 90% of the sample data is retained to prune the sample dataset, subtracting some data with lower fit degree according to the convergence parameter, and continuing to train according to the remaining sample data. After reaching the preset convergence threshold, the training of the resource allocation association analysis model is completed. For example, after the overall residual standard deviation of the pruned sample dataset is less than the preset convergence threshold, it indicates that the error fluctuation of the model becomes very small, and the model has converged and the training process ends.

[0057] Through the above method, the training of the model is completed, avoiding the difficulty of convergence caused by the project differences existing in the sample data. The core goal of model training is to infer the rationality of resource allocation through macroscopic data features such as defect density, test coverage, module relevance, and module scale features, so as to optimize the allocation strategy of test resources. However, due to the differences between different software projects, such as project scale, complexity, business requirements, etc., the resource allocation patterns in the historical reference data may not be completely consistent. These minor differences may affect the convergence and generalization ability of the model. To solve this problem, the present invention adopts residual analysis and pruning techniques to optimize the training process, helping the model to better cope with these minor differences, so as to more accurately capture the macroscopic laws in the data and ensure that the model has strong adaptability and stability among different projects. The trained resource allocation correlation analysis model is mainly used to analyze the impact of various characteristic parameters on resource allocation from a macroscopic perspective, analyze the corresponding resource allocation reference data according to the input characteristic parameters, so as to conduct risk analysis on the resource allocation strategy formulated by team members and evaluate whether there is a risk of unbalanced resource allocation.

[0058] Step S5: Conduct resource allocation risk analysis on the target test resource allocation strategy through the reference allocation strategy, and generate the risk analysis result of the target software project regarding the target test resource allocation strategy.

[0059] Specifically, through the reference allocation strategy generated by the resource allocation correlation analysis model, conduct risk analysis on the target test resource allocation strategy. For the reference allocation strategy and the target test resource allocation strategy, they respectively include the resource allocation reference proportion parameter and the resource allocation target proportion parameter of each functional module. The resource allocation deviation parameter of each functional module can be calculated through the reference allocation strategy and the target test resource allocation strategy, that is, the difference between the relevant proportion parameters of the functional module in the two strategies. Then, based on multiple resource allocation deviation parameters, conduct test resource allocation risk analysis on the target software project. It can be for multiple functional modules to calculate the overall resource allocation deviation, for example, characterized by the weighted average or sum of the resource allocation deviation parameters of all functional modules. Exemplarily, after normalizing multiple resource allocation deviation parameters, calculate the sum to obtain the resource allocation risk score of the target software project, which is used to evaluate whether the current target test resource allocation strategy has significant irrationality. If the score is high, it indicates that the gap between the current allocation strategy and the reference strategy is large, and there may be a risk of uneven test resource allocation. Taking the calculated resource allocation risk score as the risk analysis result of the target test resource allocation strategy can provide an evaluation basis for the project team on whether the test resource allocation is reasonable, help discover potential resource allocation problems, and provide guidance for subsequent resource optimization, ultimately improving the test efficiency and quality of the project.

[0060] Please refer toFigure 2 , based on the same concept of the above-mentioned risk analysis method for software project management, the present invention also provides a risk analysis system for software project management, including:

[0061] A software project data collection module, configured to collect test record data of a target software project regarding a target test phase, and a target test resource allocation strategy associated with the test record data;

[0062] A test feature extraction module, configured to extract defect density features and test coverage features respectively corresponding to multiple functional modules in the target software project from the test record data;

[0063] A function association analysis module, configured to obtain system architecture data of the target software project, perform module association analysis on multiple functional modules of the target software through the test record data and the system architecture data of the target software project, and construct a module association graph of the target software project regarding multiple functional modules;

[0064] A test resource allocation analysis module, configured to generate multiple resource allocation impact indexes for each functional module through the module association graph and the defect density features and test coverage features of multiple functional modules, perform resource allocation processing on the resource allocation impact indexes through a resource allocation association analysis model, and generate a reference allocation strategy for the target software project, where the resource allocation association analysis model is constructed based on multiple groups of historical reference test record data;

[0065] A test risk analysis module, configured to perform resource allocation risk analysis on the target test resource allocation strategy through the reference allocation strategy, and generate a risk analysis result of the target software project regarding the target test resource allocation strategy.

[0066] The above are only specific embodiments of the present invention, enabling those skilled in the art to understand or implement the present invention. Various modifications to these embodiments will be obvious to those skilled in the art. The parts not described in detail in this specification belong to the prior art well-known to those skilled in the art.

Claims

1. A risk analysis method for software project management, characterized in that, Including: Collecting test record data of a target software project regarding a target test phase, as well as a target test resource allocation strategy associated with the test record data; Extracting defect density characteristics and test coverage characteristics respectively corresponding to multiple functional modules in the target software project from the test record data; Obtaining system architecture data of the target software project, performing module association analysis on multiple functional modules of the target software through the test record data and system architecture data of the target software project, and constructing a module association graph of the target software project regarding multiple functional modules; Generating multiple resource allocation impact indices for each functional module through the module association graph and the defect density characteristics and test coverage characteristics of multiple functional modules, including calculating the module association impact index of each functional module according to multiple module dependency parameters in the module association graph, respectively performing statistical analysis on the defect density characteristics and test coverage characteristics of multiple functional modules, generating the defect distribution impact index and test coverage impact index of each functional module, and extracting the module scale impact index of the functional module from the attribute data of the functional module to obtain multiple resource allocation impact indices of the functional module; Performing resource allocation processing on the resource allocation impact indices through a resource allocation association analysis model to generate a reference allocation strategy for the target software project, where the resource allocation association analysis model is constructed based on multiple groups of historical reference test record data; Performing resource allocation risk analysis on the target test resource allocation strategy through the reference allocation strategy to generate a risk analysis result of the target software project regarding the target test resource allocation strategy.

2. The risk analysis method for software project management according to claim 1, wherein Extracting defect density characteristics and test coverage characteristics respectively corresponding to multiple functional modules in the target software project from the test record data, including: Determining the attribute data of each functional module, including code size record data and function point record data, extracting the defect quantity characteristics of each functional module from the test record data, performing defect distribution analysis on the code size record data through the defect quantity characteristics to generate the defect density characteristics of the functional module, extracting the test coverage function point quantity characteristics of each functional module from the test record data, and performing test coverage analysis on the function point record data through the test coverage function point quantity characteristics to generate the test coverage characteristics of the functional module.

3. The risk analysis method for software project management according to claim 2, characterized in that, Performing module association analysis on multiple functional modules of the target software through the test record data and system architecture data of the target software project, and constructing a module association graph of the target software project regarding multiple functional modules, including: Constructing a module association graph of the target software project regarding multiple functional modules through the system architecture data, including using multiple functional modules as nodes of the module association graph and determining the directed edges between functional modules according to the system architecture data; Separating the dependency record data between any two functional modules from the test record data, including data transmission record data and function call record data, extracting the function call parameters and data transmission parameters of each directed edge in the module association graph from multiple groups of dependency record data, and calculating the module dependency parameters of each directed edge according to the function call parameters and data transmission parameters and adding them to the module association graph.

4. A risk analysis method for software project management according to claim 3, characterized in that, Performing resource allocation risk analysis on the target test resource allocation strategy through a reference allocation strategy, and generating a risk analysis result of the target software project regarding the target test resource allocation strategy, including: Calculating the resource allocation deviation parameter of each functional module through the reference allocation strategy and the target test resource allocation strategy, performing test resource allocation risk analysis on the target software project based on multiple resource allocation deviation parameters, and calculating the resource allocation risk score of the target software project; Among them, for the reference allocation strategy and the target test resource allocation strategy, they respectively include the resource allocation reference proportion parameter and the resource allocation target proportion parameter of each functional module.

5. A risk analysis method for software project management according to claim 4, characterized in that, For the resource allocation correlation analysis model, it also includes: Constructing a training data set including multiple groups of sub-sample data through multiple groups of historical reference test record data. Each group of sub-sample data includes the historical resource allocation proportion parameter and multiple historical resource allocation influence indexes respectively corresponding to each functional module in one group of the historical reference test record data; Taking the multiple historical resource allocation influence indexes in each group of sub-sample data as the training input of the resource allocation correlation analysis model, taking the historical resource allocation proportion parameters respectively corresponding to the multiple groups of sub-sample data as the training target of the resource allocation correlation analysis model, and performing iterative training on the resource allocation correlation analysis model based on residual feature optimization through the sample data set, where the resource allocation correlation analysis model is a random forest model; Performing iterative training on the resource allocation correlation analysis model based on residual feature optimization includes calculating the residual value of each group of sub-sample data after each round of iteration, performing convergence feature analysis according to the multiple residual values of each group of sub-sample data to calculate the convergence parameter of the sub-sample data after meeting the preset number of iterations, performing sample pruning processing on the sample data set based on the preset pruning parameter and the convergence parameter of the sub-sample data, and completing the training of the resource allocation correlation analysis model after reaching the preset convergence threshold.

6. A risk analysis system for software project management, characterized in that, A risk analysis method for software project management used to execute any one of claims 1 - 5, including: A software project data collection module, used to collect the test record data of the target software project regarding the target test phase, and the target test resource allocation strategy associated with the test record data; A test feature extraction module, used to extract the defect density feature and test coverage feature respectively corresponding to multiple functional modules in the target software project from the test record data; A functional correlation analysis module, used to obtain the system architecture data of the target software project, perform module correlation analysis on multiple functional modules of the target software through the test record data and the system architecture data of the target software project, and construct a module correlation graph of the target software project regarding multiple functional modules; A test resource allocation analysis module, used to generate multiple resource allocation influence indexes of each functional module through the module correlation graph and the defect density feature and test coverage feature of multiple functional modules, perform resource allocation processing on the resource allocation influence indexes through the resource allocation correlation analysis model, and generate the reference allocation strategy of the target software project, where the resource allocation correlation analysis model is constructed based on multiple groups of historical reference test record data; A test risk analysis module is used to perform resource allocation risk analysis on a target test resource allocation strategy by referring to an allocation policy, and generate a risk analysis result of a target software project regarding the target test resource allocation strategy.

Images