A prediction method and system reflecting changes in personalized needs of business personnel

Abstract

The present invention relates to a prediction method and system that reflects the personalized demand changes of business personnel. By analyzing the historical behaviors of business personnel and extracting effective information, the functional modules that may have demand changes in the project and the number of demand changes are predicted, and the prediction results are provided to the project developers and testers. For different projects and different business personnel, the functional modules that may have demand changes in the project and the number of possible changes are reasonably and accurately predicted, which can effectively guide the developers and testers to prepare relevant response measures for demand changes, so as to achieve the purpose of ensuring project quality.

Description

Technical Field

[0001] The present invention relates to the technical field of computer software development and testing, and in particular to a prediction method and system for reflecting changes in personalized requirements of business personnel. Background Art

[0002] Requirements changes, i.e. changes to project or software development requirements, refer to functional or non-functional changes to a project or software proposed by business personnel after completing the software requirements document and before the software is delivered. Requirements changes will inevitably affect the developers and testers of the project. After the requirements change, the development workload and testing workload of the project will increase, but the project cycle will not necessarily be extended. Therefore, requirements changes have an important impact on the development quality and testing quality of the project. However, before each change in the project, developers and testers are not informed and cannot prepare relevant countermeasures for the requirements changes.

[0003] If it is possible to predict the changes in project requirements, especially to obtain the functional modules that require specific changes and the number of possible changes, it can support developers and testers to prepare corresponding countermeasures in advance, which can greatly improve the efficiency of development and testing work.

[0004] At present, there are some methods that use related algorithms to predict demand changes, such as the CRF-LSTM algorithm and the Bi-LSTM algorithm. However, these methods often do not take into account the attributes of the person who proposes the demand (that is, the business person), resulting in the prediction results being inaccurate and having limited applicability. Summary of the invention

[0005] In order to address the deficiencies in the prior art, the present invention proposes a prediction method and system that reflects the personalized demand changes of business personnel. By analyzing the historical behaviors of business personnel and extracting effective information, the functional modules that may have demand changes in the project and the number of demand changes are predicted, and the prediction results are provided to the project developers and testers. For different projects and different business personnel, the functional modules that may have demand changes in the project and the number of possible changes can be reasonably and accurately predicted, which can effectively guide developers and testers to prepare relevant response measures for demand changes, thereby achieving the purpose of ensuring project quality.

[0006] To achieve the above objectives, the technical solutions adopted by the present invention include:

[0007] A prediction method that reflects changes in personalized needs of business personnel, characterized by comprising:

[0008] S1. Integrate the test assets to establish an asset library, wherein the asset library at least includes historical behavior records of business personnel operating test assets, and the historical behavior records at least include the test assets corresponding to the behavior, the behavior category, and the project to which the behavior belongs;

[0009] S2. Determine the target project to be analyzed, and retrieve the target historical behavior records belonging to the target project from the asset library according to the target project;

[0010] S3. According to the historical behavior records of the target, quantitative calculation is performed to obtain a set of preference values ​​of each business personnel for each test asset;

[0011] S4. Determine the functional modules included in the target project, and quantitatively calculate and obtain a set of correlations between each test asset and each functional module included in the target historical behavior record;

[0012] S5. Taking business personnel as groups, using preference values ​​and correlation calculation to obtain a set of demand impact factors corresponding to specific business personnel, wherein the set of demand impact factors includes demand impact factors between each test asset and each functional module, and the use of preference values ​​and correlation calculation includes multiplying the preference value of the specific business personnel for a specific test asset and the correlation between the test asset and the specific functional module to obtain the demand impact factor between the specific test asset corresponding to the specific business personnel and the specific functional module;

[0013] S6. Introduce a set of demand influencing factors to establish a calculation model for predicting demand changes, and generate a demand change prediction model corresponding to a specific business person;

[0014] S7, determining the functional module of the project to be predicted and retrieving the historical behavior record to be predicted belonging to the project to be predicted, and calculating and obtaining a set of demand influencing factors corresponding to the project to be predicted;

[0015] S8. Input the set of demand influencing factors of the project to be predicted into the demand change prediction model to obtain the demand change prediction result of the project to be predicted.

[0016] Furthermore, the test assets include any one or more combinations of test case libraries, test case writing elements, function points of the system under test, test rule points, automation case libraries, business journeys, test data construction, test-related technologies, and historical production events; and the behavior categories include search and browse.

[0017] Furthermore, the step S3 includes the following sub-steps:

[0018] S31, extracting the number of searches i and the number of views j of a specific business person on a specific test asset from the target historical behavior record;

[0019] S32. Use Formula 1 to calculate the preference value z of a specific business person for a specific test asset.

[0020] z=i+2*j Formula 1.

[0021] Furthermore, the step S4 includes the following sub-steps:

[0022] S41, segmenting the test assets and the functional modules respectively and matching them to obtain the test asset word vector a and the functional module word vector b, wherein the test asset word vector a and the functional module word vector b have the same vector dimension;

[0023] S42, using formula 2 to calculate a set of associations y between specific test assets and specific functional modules,

[0024]

[0025] Among them, i is the corresponding vector feature of the test asset word vector a and the function module word vector b, and n is the total vector dimension of the test asset word vector a and the function module word vector b.

[0026] Furthermore, the demand change prediction model is a deep neural network model.

[0027] Furthermore, the input value of the deep neural network model includes the demand influencing factor, and the output value includes the demand change probability and the number of demand changes, the demand change probability has a value range of [0,1], and the number of demand changes is a non-negative integer.

[0028] The present invention also relates to a prediction system that reflects changes in personalized needs of business personnel, characterized in that it includes:

[0029] The asset library establishment module is used to integrate test assets to establish an asset library;

[0030] A preference value calculation module is used to obtain a set of preference values ​​by quantitative calculation;

[0031] A correlation calculation module is used to obtain a set of correlations through quantitative calculation;

[0032] A demand impact factor calculation module is used to calculate and obtain a set of demand impact factors;

[0033] The prediction module is used to generate a demand change prediction model and use the demand change prediction model to perform prediction to obtain the demand change prediction result.

[0034] The present invention also relates to a computer-readable storage medium, characterized in that a computer program is stored on the storage medium, and the computer program implements the above method when executed by a processor.

[0035] The present invention also relates to an electronic device, characterized in that it comprises a processor and a memory;

[0036] The memory is used to store the asset library;

[0037] The processor is used to execute the above method by calling the asset library.

[0038] The beneficial effects of the present invention are:

[0039] The prediction method and system for reflecting the personalized demand changes of business personnel described in the present invention are adopted. By analyzing the historical behavior records of business personnel on relevant test assets in the asset library, the preference values ​​of business personnel corresponding to the test assets are generated, and the demand influencing factors are generated by combining the correlation between the test assets and the projects. The demand change prediction model takes into account both the objective factors of the project and the subjective factors of the business personnel, and is closer to the actual development process of the project. It can reasonably and accurately predict the functional modules of the project demand changes and the possible number of changes for different projects and different business personnel, and effectively guide developers and testers to prepare relevant response measures for demand changes, so as to achieve the purpose of ensuring the quality of project development. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 The figure is a flow chart of a method for predicting changes in personalized needs of business personnel according to the present invention.

[0041] Figure 2 The present invention is a schematic diagram of the prediction system structure that reflects the personalized demand changes of business personnel. DETAILED DESCRIPTION

[0042] In order to more clearly understand the content of the present invention, it will be described in detail with reference to the drawings and embodiments.

[0043] A first aspect of the present invention relates to a Figure 1 The process shown in the figure reflects the prediction method of the personalized demand change of business personnel, including:

[0044] S1. Integrate test assets to establish an asset library, wherein the test assets include any one or more combinations of a test case library, test case writing elements, function points of the system under test, test rule points, an automation case library, a business journey, a test data structure, test-related technologies, and historical production events. The asset library includes at least a historical behavior record of a business person operating a test asset, and the historical behavior record includes at least a test asset corresponding to the behavior, a behavior category (including search and browse), and a project to which the behavior belongs;

[0045] S2. Determine the target project to be analyzed, and retrieve the target historical behavior records belonging to the target project from the asset library according to the target project;

[0046] S3. According to the historical behavior records of the target, quantitative calculation is performed to obtain the set of preference values ​​of each business personnel for each test asset, including:

[0047] S31, extracting the number of searches i and the number of views j of a specific business person on a specific test asset from the target historical behavior record;

[0048] S32. Use Formula 1 to calculate the preference value z of a specific business person for a specific test asset.

[0049] z=i+2*j Formula 1;

[0050] S4. Determine the functional modules included in the target project, and quantitatively calculate and obtain the set of correlations between each test asset and each functional module included in the target historical behavior record, specifically including:

[0051] S41, segmenting the test assets and the functional modules respectively and matching them to obtain the test asset word vector a and the functional module word vector b, wherein the test asset word vector a and the functional module word vector b have the same vector dimension;

[0052] S42, using formula 2 to calculate a set of associations y between specific test assets and specific functional modules,

[0053]

[0054] Where i is the corresponding vector feature of the test asset word vector a and the function module word vector b, and n is the total vector dimension of the test asset word vector a and the function module word vector b;

[0055] S5. Taking business personnel as groups, using preference values ​​and correlation calculation to obtain a set of demand impact factors corresponding to specific business personnel, wherein the set of demand impact factors includes demand impact factors between each test asset and each functional module, and the use of preference values ​​and correlation calculation includes multiplying the preference value of the specific business personnel for a specific test asset and the correlation between the test asset and the specific functional module to obtain the demand impact factor between the specific test asset corresponding to the specific business personnel and the specific functional module;

[0056] S6. A deep neural network model for predicting demand changes is established by introducing a set of demand influencing factors, and a demand change prediction model corresponding to a specific business personnel is obtained, wherein the input value includes the demand influencing factor, and the output value includes the probability of demand change and the number of demand changes, wherein the value range of the probability of demand change is [0, 1], and the number of demand changes is a non-negative integer;

[0057] S7, determining the functional module of the project to be predicted and retrieving the historical behavior record to be predicted belonging to the project to be predicted, and calculating and obtaining a set of demand influencing factors corresponding to the project to be predicted;

[0058] S8. Input the set of demand influencing factors of the project to be predicted into the demand change prediction model to obtain the demand change prediction result of the project to be predicted.

[0059] The prerequisite for implementing the above method is to have an asset library that can be searched and browsed by business personnel. By recording the search and browsing operations of business personnel, data is formed to judge the preferences of business personnel, and the preferences are quantitatively introduced into the prediction and calculation process of demand changes.

[0060] In the specific implementation process, assuming that the number of searches for asset A by the business personnel is i, the number of views of asset A is j, and the preference value of the business personnel for the relevant test assets is z, the formula 1 is specified: z = i + 2*j, thereby obtaining the preference value z of the business personnel for asset A. By repeating this process, the set z of the preference values ​​of the business personnel for all relevant test assets can be obtained. j , where j represents a different test asset.

[0061] At the same time, by quantifying the correlation between test assets and functional modules involved in the project, we can obtain a collection of correlations. Assuming that business personnel searched or browsed three test assets A1, A2, and A3 during the project cycle, with preference values ​​of z1, z2, and z3 respectively, and the project involves three functional modules B1, B2, and B3, then we need to consider the full coverage combination of correlations between test assets and functional modules, that is, the correlation between A1 and B1, the correlation between A1 and B2, the correlation between A1 and B3, the correlation between A2 and B1, the correlation between A2 and B2, the correlation between A2 and B3, the correlation between A3 and B1, the correlation between A3 and B2, and the correlation between A3 and B3.

[0062] Preferably, the word vector cosine algorithm shown in the following formula 2 is used to calculate the relevance:

[0063]

[0064] Taking the correlation between A1 and B1 as an example, A1 and B1 are first segmented, each segmented word represents a dimension of the space, and then converted into word vectors a and b in the multidimensional space. The similarity value is calculated using Formula 2. The similarity value is between 0 and 1. The higher the value, the more similar it is.

[0065] Through the above steps, we get the preference value z and the correlation degree y between the test asset and the project functional module. We multiply the preference value and the correlation value to get the demand impact factor u. For example, the demand impact factor u between the test asset A1 and the functional module B1 is 11 =z1×y 11 .

[0066] After obtaining the required demand influencing factors, it is preferred to use a deep learning algorithm to establish a model. The demand influencing factors can be used as parameter indicators that affect the prediction results and as input layer neurons of the deep neural network model. The model is trained using historical data to obtain the optimal solution of the deep neural network model.

[0067] For example, if a deep learning model is established for functional module B1 in the project, the association values ​​between test assets A1, A2, A3 and functional module B1 are y 11 ,y 21 ,y 31 , so the demand influencing factor is u 11 、u 21 、u 31 .

[0068] The above three demand influencing factors are the three input layer neurons of the deep neural network model, and the model is set to have two output values, namely w and v, where w is the probability of demand change, with a value range of [0,1], and v is the number of demand changes, with a value range of non-negative integers.

[0069] The input neurons of the neural network model are called the input layer, corresponding to the demand impact factor u 11 、u 21 、u 31 ; The hidden layer of the neural network model, including the number of hidden layers m and the number of nodes n in each hidden layer, represented by H mn represents the nth node of the mth hidden layer, and the hidden layer is followed by the original output layer, corresponding to two output values ​​w and v. Neurons in both the hidden layer and the original output layer need to be delinearized using an activation function, and the preferred activation function used in the present invention is the sigmoid function: The softmax layer is an additional processing layer that converts the output of the neural network into a probability distribution. Suppose the output of the original output layer is x1, x2, ..., x n , then the output after softmax regression processing is The final output layer corresponds to two output values ​​w and v.

[0070] The historical data of the system corresponding to the model is collected as a training set. After the number of hidden layers is determined, the deep neural network model is iteratively trained using the error back propagation algorithm (BP algorithm). After the training converges, the weight of the neural network model can be obtained, that is, the coefficient corresponding to the connection between nodes. After the model training is completed, the demand influencing factors between the functional modules and various test assets in the collection project are used as input to obtain output, thereby predicting the probability and number of demand changes.

[0071] The present invention also relates to a structure such as Figure 2 The forecasting system that reflects the personalized demand changes of business personnel shown in the figure includes:

[0072] The asset library establishment module is used to integrate test assets to establish an asset library;

[0073] A preference value calculation module is used to obtain a set of preference values ​​by quantitative calculation;

[0074] A correlation calculation module is used to obtain a set of correlations through quantitative calculation;

[0075] A demand impact factor calculation module is used to calculate and obtain a set of demand impact factors;

[0076] The prediction module is used to generate a demand change prediction model and use the demand change prediction model to perform prediction to obtain the demand change prediction result.

[0077] By using this system, the above method can be implemented and demand change prediction results can be obtained.

[0078] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A prediction method that reflects changes in personalized needs of business personnel, characterized in that: include: S1. Integrate the test assets to establish an asset library, wherein the asset library at least includes historical behavior records of business personnel operating test assets, and the historical behavior records at least include the test assets corresponding to the behavior, the behavior category, and the project to which the behavior belongs; S2. Determine the target project to be analyzed, and retrieve the target historical behavior records belonging to the target project from the asset library according to the target project; S3. According to the historical behavior records of the target, quantitative calculation is performed to obtain a set of preference values ​​of each business personnel for each test asset; S4. Determine the functional modules included in the target project, and quantitatively calculate and obtain a set of correlations between each test asset and each functional module included in the target historical behavior record; S5. Taking business personnel as groups, using preference values ​​and correlation calculation to obtain a set of demand impact factors corresponding to specific business personnel, wherein the set of demand impact factors includes demand impact factors between each test asset and each functional module, and the use of preference values ​​and correlation calculation includes multiplying the preference value of the specific business personnel for a specific test asset and the correlation between the test asset and the specific functional module to obtain the demand impact factor between the specific test asset corresponding to the specific business personnel and the specific functional module; S6. Introduce a set of demand influencing factors to establish a calculation model for predicting demand changes, and generate a demand change prediction model corresponding to a specific business person; S7, determining the functional module of the project to be predicted and retrieving the historical behavior record to be predicted belonging to the project to be predicted, and calculating and obtaining a set of demand influencing factors corresponding to the project to be predicted; S8, inputting a set of demand influencing factors of the project to be predicted into a demand change prediction model to obtain a demand change prediction result of the project to be predicted; Wherein, the step S3 includes the following sub-steps: S31, extracting the number of searches i and the number of views j of a specific business person on a specific test asset from the target historical behavior record; S32, calculating and obtaining the preference value z of a specific business person for a specific test asset, ; The step S4 comprises the following sub-steps: S41, segmenting the test assets and the function modules respectively and matching them to obtain the test asset word vector a and the function module word vector b, wherein the test asset word vector a and the function module word vector b have the same vector dimension; S42, calculating and obtaining a set of associations y between specific test assets and specific functional modules, ; Among them, i is the corresponding vector feature of the test asset word vector a and the function module word vector b, and n is the total vector dimension of the test asset word vector a and the function module word vector b.

2. The method according to claim 1, characterized in that The test assets include any one or more combinations of test case libraries, test case writing elements, function points of the system under test, test rule points, automation case libraries, business journeys, test data construction, test-related technologies, and historical production events; the behavior categories include search and browse.

3. The method according to claim 1, characterized in that The demand change prediction model is a deep neural network model.

4. The method according to claim 3, characterized in that The input value of the deep neural network model includes the demand influencing factor, and the output value includes the demand change probability and the number of demand changes. The demand change probability has a value range of [0,1], and the number of demand changes is a non-negative integer.

5. A prediction system that reflects the personalized demand changes of business personnel, which is used to implement the method according to any one of claims 1 to 4 when executed, characterized in that: include: The asset library establishment module is used to integrate test assets to establish an asset library; A preference value calculation module is used to obtain a set of preference values ​​by quantitative calculation; A correlation calculation module is used to obtain a set of correlations through quantitative calculation; A demand impact factor calculation module is used to calculate and obtain a set of demand impact factors; The prediction module is used to generate a demand change prediction model and use the demand change prediction model to perform prediction to obtain the demand change prediction result.

6. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 4 is implemented.

7. An electronic device, characterized in that: including a processor and a memory; The memory is used to store the asset library; The processor is used to execute the method according to any one of claims 1 to 4 by calling the asset library.

Images