An AI-based intelligent analysis and prediction method and system for software requirements

Abstract

The present invention relates to the technical field of intelligent demand analysis of AI, and discloses a method and system for intelligent analysis and prediction of software requirements based on AI, including: collecting usage data of the software when the software product is in use; and generating demand feedback information by obtaining user feedback; using a transfer learning strategy to perform demand analysis on the usage data and demand feedback information; performing a game between the result of value analysis and the resource occupancy to generate a final software requirement prediction result. While ensuring the demand satisfaction degree, it minimizes resource consumption and improves development efficiency. By constructing a functional value evaluation system, the present invention can balance the functional benefits and implementation costs, provide a scientific and automated decision-making basis for enterprises, greatly reduce the software iteration cycle, and improve the user experience.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent analysis of AI requirements, and specifically to a method and system for intelligent analysis and prediction of software requirements based on AI. Background Art

[0002] In the modern software development process, accurately identifying and meeting user requirements has become the key to enhancing product competitiveness. However, traditional software requirement analysis methods usually rely on user interviews, questionnaires, market analysis, etc. These methods not only have high costs and long cycles, but also have problems such as strong subjectivity and untimely data updates. In addition, the diversity and uncertainty of user requirements make it difficult for development teams to accurately judge which functions have high value and which functions may be ignored or waste development resources.

[0003] In recent years, with the development of big data, artificial intelligence (AI) and machine learning (ML) technologies, software requirement analysis has gradually evolved towards data-driven and intelligent directions. By analyzing user behavior data, it is possible to more deeply explore users' operation habits, function preferences, and potential unmet requirements during software use. For example, based on information such as users' operation logs, clickstream data, and search query records, an intelligent requirement analysis system can be constructed to more accurately understand the changing trends of user requirements.

[0004] In addition, cross-domain data migration has also attracted wide attention in software requirement analysis. There may be similarities in interface structures, interaction logics, and function modules between different software products. Therefore, through data migration and learning, the dependence of new systems on large-scale user feedback data can be reduced, and the efficiency of requirement prediction can be improved. At the same time, how to balance function value and resource consumption during the requirement analysis process, optimize resource allocation, and enable software development to take into account user experience, development costs, and system performance has become an important research direction in the industry. Summary of the Invention

[0005] In view of the above existing problems, the present invention is proposed.

[0006] Therefore, the technical problems solved by the present invention are: existing software requirement intelligent analysis methods have problems such as the analysis results being unable to adapt to the deployment of software and how to select high-quality requirements.

[0007] To solve the above technical problems, the present invention provides the following technical solution: A method for intelligent analysis and prediction of software requirements based on AI, including:

[0008] When the software product is in use, collect the usage data of the software; and generate requirement feedback information by obtaining user feedback;

[0009] Using a transfer learning strategy, perform requirement analysis on the usage data and requirement feedback information;

[0010] Perform a game between the results of value analysis and the resource occupancy to generate the final software requirement prediction result.

[0011] As a preferred solution of the AI-based intelligent software requirement analysis and prediction method of the present invention, wherein: the usage data includes the operation records of users in the software, the query records of the search box, and the records of users using multiple software functions of the software at the same time;

[0012] Among them, in each record, the time when the record occurs is included.

[0013] As a preferred solution of the AI-based intelligent software requirement analysis and prediction method of the present invention, wherein: the requirement feedback information includes the text information provided by users through the feedback channel regarding the software requirements.

[0014] As a preferred solution of the AI-based intelligent software requirement analysis and prediction method of the present invention, wherein: the transfer learning strategy includes obtaining different software products. Let the i-th software product be ; starting from the main menu as the starting node, using each interactive sub-menu as a node, using each software function as the end point, and using the connection line to represent the interactive relationship; respectively construct the paths for the implementation of each function of this software and as the function graph;

[0015] During the process of user interaction behavior, when the operation is exited at a node other than the end point and the starting point, it is recorded as a hidden event; count the frequency of occurrence of the hidden event: ; wherein, represents the frequency of occurrence of the hidden event at node b; represents the number of times the hidden event occurs at node b; represents the total number of times of reaching node b through interaction;

[0016] Through a pre-trained deep learning algorithm, predict the requirements of this software ; Input: and the function graphs of, and the frequency of occurrence of each hidden event; Output: the end points that are in and not in at the same time, and at the same time output the probability of each end point representing the software requirement to obtain the end point set Z.

[0017] As a preferred solution of the AI-based intelligent software requirement analysis and prediction method of the present invention, wherein: according to the requirement feedback information, for Perform secondary matching on the software functions in it, and based on the results of the secondary matching, conduct value analysis on each endpoint output by the deep learning algorithm; the secondary matching includes, after extracting keywords from the demand feedback information, calculating the distance from each endpoint in the endpoint set Z; for each endpoint, multiply the similarity by the number of each keyword and then sum to obtain the value of each endpoint in the endpoint set Z, and each endpoint represents a demand selection;

[0018] Suppose the user uses m functions at the same time, expressed as ; randomly combine the m software, and calculate the value of each function combination: ; where, represents the value of the y-th combination; represents the probability that when the function represented by the n-th element in the combination is in use, the functions represented by the elements in the y-th combination are in use at the same time; n represents the number of elements in the y-th combination; represents the expansion coefficient, ; each function combination represents a function selection.

[0019] As a preferred solution of the AI-based intelligent analysis and prediction method for software requirements of the present invention, wherein: the occupation of resources includes, for each function selection, on the basis of the existing functions of this software to be implemented, resources need to be occupied for calculation;

[0020] Through a pre-trained machine learning algorithm, calculate the resource occupation of each function selection; input: function selection and the existing functions; output: the resources that need to be occupied by the function selection.

[0021] As a preferred solution of the AI-based intelligent analysis and prediction method for software requirements of the present invention, wherein: the game between the results of value analysis and the occupation of resources includes, obtaining all function selections, set as G; randomly extracting from the G function selections, the number of each extraction is at most G, and the function selection combinations extracted each time are different;

[0022] Obtain Q function selection combinations, calculate the total value and total occupied resources of each function selection combination; after removing the function selection combinations whose total occupied resources exceed λ, obtain the alternative function selection combinations; where, λ represents the preset total occupied resource threshold;

[0023] Use the alternative function selection combinations as the prediction results, and at the same time screen and mark high-quality combinations to generate an analysis report;

[0024] The high-quality combination includes, for each prediction result, conducting a game with the goal of maximizing the total value and minimizing the total occupied resources; establishing a two-dimensional coordinate system where the abscissa represents the total occupied resources and the ordinate represents the total value;

[0025] Multiply the total value of each prediction result by and use it as the ordinate; use the total occupied resources of each prediction result as the abscissa; deploy each prediction result in the two-dimensional coordinate system;

[0026] where = the maximum value of the total value in the prediction results / to make the maximum values of both the abscissa and ordinate be If a prediction result falls within the triangle formed by the three points , , , then it is judged as a high-quality combination;

[0027] In the analysis report, record each prediction result, as well as the value analysis result and the occupied resources of the prediction result.

[0028] An AI-based intelligent software requirement analysis and prediction system adopting any method described in the present invention, wherein:

[0029] A collection unit, when the software product is in use, collects the usage data of the software; and generates requirement feedback information by obtaining user feedback;

[0030] An analysis unit, using a transfer learning strategy, conducts requirement analysis on the usage data and requirement feedback information;

[0031] A game unit, conducts a game between the result of value analysis and the resource occupation to generate the final software requirement prediction result.

[0032] A computer device, including: a memory and a processor; the memory stores a computer program, wherein: when the processor executes the computer program, it implements the steps of any method described in the present invention.

[0033] A computer-readable storage medium, on which a computer program is stored, wherein: when the computer program is executed by a processor, it implements the steps of any method described in the present invention.

[0034] Advantages of the present invention: The intelligent analysis and prediction method for software requirements based on AI provided by the present invention can effectively improve the accuracy and optimization ability of software requirement prediction through intelligent requirement analysis, transfer learning, resource optimization, and game theory decision-making. Compared with traditional methods, the present invention can automatically extract user behavior data, combine deep learning models to mine potential requirements, and use transfer learning strategies to predict possible function missing points across software. In addition, the present invention combines game theory optimization algorithms to minimize resource consumption while ensuring requirement satisfaction, thereby improving development efficiency. By constructing a functional value evaluation system, the present invention can balance functional benefits and implementation costs, provide scientific and automated decision-making basis for enterprises, significantly reduce the software iteration cycle, and improve user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention, and those of ordinary skill in the art can obtain other drawings based on these drawings without creative efforts.

[0036] Figure 1 It is the overall flowchart of an intelligent analysis and prediction method for software requirements based on AI provided by the first embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0037] In order to make the above objects, features, and advantages of the present invention more obvious and understandable, the specific embodiments of the present invention will be described in detail below with reference to the drawings of the specification. Obviously, the described embodiments are some embodiments of the present invention, rather than all embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0038] Embodiment 1, referring to Figure 1 , which is an embodiment of the present invention, provides an intelligent analysis and prediction method for software requirements based on AI, including:

[0039] S1: When the software product is in use, collect the usage data of the software; and generate requirement feedback information by obtaining user feedback.

[0040] The usage data includes the operation records of users in the software, the query records in the search box, and the records of users using multiple software functions simultaneously; among them, each record contains the time when the record occurred. The operation records include the sequence of human-computer interaction generated during a single continuous operation. The query records include the text content to be queried entered in the search box. When users use multiple software functions simultaneously, not only the time is collected, but also what software functions are used simultaneously. The requirement feedback information includes the text information provided by users through the feedback channel regarding the requirements for the software.

[0041] That is to say, in fact, the feedback includes two types. One is the implicit feedback of "collecting the usage data of the software", through which the implicit expression of requirements can be predicted. The second is the explicit feedback directly provided by users. Through the combined action of the explicit feedback and the implicit feedback, the subsequent analysis and prediction processes are realized. The explicit feedback can directly guide the software optimization direction, while the implicit feedback can supplement the requirements that users fail to clearly express. The combination of the two can more comprehensively construct the user requirement model and improve the accuracy of subsequent requirement prediction.

[0042] S2: Use the transfer learning strategy to perform requirement analysis on the usage data and requirement feedback information.

[0043] The transfer learning strategy includes obtaining different software products. Let the i-th software product be ; starting from the main menu as the starting node, taking each interactive sub-menu as a node, and each software function as the end point, with the connection lines representing the interactive relationships; respectively construct the paths for the implementation of each function of this software and as the function graph. By constructing the paths, there is no need to vectorize the text data, nor is there a need to perform complex relationship analysis. The analysis can be completed through the simplest comparison relationship, which can greatly reduce the consumption of computing resources.

[0044] During the process of users' interaction behavior, when exiting the operation at nodes other than the end point and the starting point, it is recorded as an implicit event; count the frequency of the occurrence of the implicit event: ; among them, represents the frequency of the occurrence of the implicit event at node b; represents the number of times the implicit event occurs at node b; represents the total number of times of reaching node b through interaction. Through a deep learning algorithm with pre-training (training set: validation set = 8:2), predict the requirements of this software ; input: and The functional map and the frequency of each implicit event (the main reason for inputting this element is to increase the attention to implicit events. Because when selecting functions, users mostly search in similar sub-menus. An implicit event can generally be understood as "not finding a certain function" and may thus be needed); Output: and only has endpoints that are not in, and at the same time output the probability that each endpoint represents a software requirement (that is to say, this deep learning algorithm judges each endpoint to calculate the probability that it is a "requirement"), obtaining the endpoint set Z.

[0045] It should be noted that during the software operation process, users usually enter the function page from the main menu. However, if they frequently exit in the lower-level menu, it may mean that the users fail to find the expected function, or the current menu organization does not meet the users' expectations. By counting the frequency of implicit events (that is, the number of times users exit at a certain menu point), it is possible to quantify whether there are missing functions with high demand but unmet in a certain interface. By constructing a functional map and analyzing the hierarchical relationship between different functions, it is ensured that the requirements analysis not only considers single interface operations but also the relevance between different functions. For example, if users frequently exit in the "Privacy Options" of the "Settings" menu, it is very likely that this menu lacks specific functions (such as "Advanced Privacy Settings"). Since users mostly search for functions in similar sub-menus in most cases, the deep learning algorithm should pay attention to those high-frequency implicit events and increase the weight of these data in demand prediction. This model not only analyzes the usage of existing functions but can also infer functions that are not available in the existing software but may be needed, and calculate the probability of each function as a requirement, thus generating an optimal demand prediction.

[0046] In this embodiment, the core task of the deep learning algorithm is to predict the functions that may be missing in the software based on the functional map and user implicit events, and calculate their demand probabilities. To achieve this goal, the deep learning algorithm used is the Graph Neural Network (GNN). The functional map is a natural graph structure. There are hierarchical relationships and interaction paths between the menus and functions of the software, which can be modeled by a graph. Each menu item and function is a node, and the click / jump behavior is an edge. GNN is very suitable for processing this kind of graph-structured data, and can learn the associations between different functions to help infer functions that are not provided but may be needed. Since the functional layouts of different software may have similarities, GNN can use the functional maps of different software as training data to improve the prediction ability through transfer learning.

[0047] In other alternative embodiments, a Transformer + self-attention mechanism can also be used. The Transformer model can process sequential data and calculate the interaction relationship between the user among different interfaces and functions through the self-attention mechanism. Or LSTM (Long Short-Term Memory network); LSTM can remember long-term historical operations, is suitable for modeling the user's interaction behavior sequence, and is especially suitable for scenarios with time-order characteristics of operations.

[0048] According to the said demand feedback information, perform secondary matching on the software functions in, and perform value analysis on each endpoint output by the deep learning algorithm according to the results of the secondary matching; the secondary matching includes, after extracting keywords from the demand feedback information, calculating the distance from each endpoint in the endpoint set Z; for each endpoint, multiply the similarity by the number of each keyword and then perform a summation process to obtain the value of each endpoint in the endpoint set Z, and each endpoint represents a demand selection.

[0049]

[0050] Among them, T represents the number of endpoints in the endpoint set Z; t represents the index, represents the value of the t-th endpoint, represents the similarity between the t-th endpoint and the keyword v; represents the number of the keyword v; V represents the total number of keywords.

[0051] It should be noted that generally, the user's feedback will not be directly used as the basis for demand generation, but only for reference. Because some user feedback may be rather absurd and impossible to implement. However, the functions represented by the "endpoints" after transfer learning can be implemented. Therefore, the user's feedback only serves as a reference factor for value in this solution.

[0052] It should be mentioned that in demand analysis, the text information feedback by users is often unstructured. In order to accurately match the user's needs, through secondary matching, it is ensured that the final demand prediction result highly coincides with the user's actual needs, reducing the interference of irrelevant or low-value functions.

[0053] Suppose the user simultaneously uses m functions, which is expressed as ; randomly combine the m software and calculate the value of each function combination: ; among them, represents the value of the y-th combination; represents the probability that the functions represented by the elements in the y-th combination are used simultaneously when the function represented by the n-th element in the combination is used; n represents the number of elements in the y-th combination; Represents the expansion coefficient, (The value of the combination is much greater than "the value of each end point in the end point set Z". Since this part of the combination of existing functions hardly occupies resources or occupies very few resources, the cost performance is relatively high); each function combination represents a function selection.

[0054] S3: Conduct a game between the results of value analysis and resource occupation to generate the final software requirement prediction result.

[0055] Furthermore, the occupation of resources includes, for each function selection, on the basis of the existing functions of this software To be implemented, resource calculation is required. Through a pre-trained machine learning algorithm, calculate the resource occupation of each function selection; Input: function selection and The existing functions; Output: The resources that the function selection needs to occupy.

[0056] Through a pre-trained machine learning model, combined with the existing functions of the software, predict the resource occupation such as storage, calculation, bandwidth, and development time of each function selection, ensure that the optimization plan maximizes the value while achieving the optimal utilization of resources, so as to balance user requirements and development feasibility, and improve the software iteration efficiency and cost performance.

[0057] In this embodiment, the gradient boosting decision tree (GBDT) is used for resource occupation prediction for the following reasons: The input features are non-linearly complex, and GBDT can handle non-linear mapping: Resource occupation is affected by function call chains, UI complexity, data processing requirements, etc., and GBDT is suitable for modeling such complex relationships. High training efficiency and low consumption of computing resources: Compared with deep learning, GBDT has a faster training speed and occupies less resources, and is suitable for prediction tasks of medium-scale data sets. Strong interpretability and support for feature importance analysis: It can analyze which factors have the greatest impact on resource occupation and help developers optimize resource allocation strategies.

[0058] Furthermore, obtain all function selections, set as G; randomly select from the G function selections, and the number of each selection is at most G, and the function selection combinations for each selection are different. Obtain Q function selection combinations, calculate the total value and total occupied resources of each function selection combination; after removing the function selection combinations whose total occupied resources exceed λ, obtain the alternative function selection combinations; where λ represents the preset total occupied resource threshold.

[0059] Use the alternative function selection combinations as the prediction results, and at the same time screen and mark high-quality combinations to generate an analysis report.

[0060] The high-quality combination includes, for each prediction result, conducting a game with the goal of maximizing the total value and minimizing the total occupied resources; establishing a two-dimensional coordinate system where the abscissa represents the total occupied resources and the ordinate represents the total value. Multiply the total value of each prediction result by and use it as the ordinate; use the total occupied resources of each prediction result as the abscissa; deploy each prediction result in the two-dimensional coordinate system.

[0061] Among them, = the maximum value of the total value in the prediction results / , so that the maximum values of the abscissa and ordinate are both . If the prediction result falls into the triangle formed by , , these three points, it is determined as a high-quality combination.

[0062] In the analysis report, record each prediction result, as well as the value analysis result and the occupied resources of the prediction result. It should be noted that the goal of the above solution is to optimize the selection of function combinations among multiple function selection schemes, ensuring that the results of software requirement prediction are both highly valuable and meet resource constraints. The specific purposes include: balancing function value and resource occupancy, and improving software optimization efficiency: by calculating the total value and resource occupancy of function combinations, screening function sets with high returns and low costs, and avoiding waste of development resources. Set a resource threshold λ to eliminate function combinations that occupy too many resources, ensuring that the optimization plan meets the actual development feasibility. Use game theory optimization to find the optimal function combination: adopt a dual-objective optimization strategy of maximizing the total value & minimizing resources, ensuring that the final plan meets user needs and does not cause a system burden. Through the two-dimensional coordinate game strategy, screen among different function combinations, making the final recommended plan optimal in terms of value-resource ratio. Screen high-quality combinations through geometric methods to improve decision-making efficiency: in the two-dimensional coordinate system, the abscissa represents resource occupancy and the ordinate represents value, constructing a high-quality combination area (a triangle formed by three key points). If the function combination falls into this area, it is considered that the combination has optimality in the trade-off between resources and value, ensuring that the selected plan is the most cost-effective for software optimization. Automatically generate an analysis report to assist software function decision-making: visualize the optimization results to help developers more intuitively understand the advantages and disadvantages of different plans, reducing the subjective error of manual judgment. Automatically mark "high-quality combinations" to improve the scientific, intelligent level and development efficiency of software product optimization.

[0063] On the other hand, this embodiment also provides an AI-based intelligent analysis and prediction system for software requirements, which includes:

[0064] A collection unit that, when the software product is in use, collects the usage data of the software; and generates demand feedback information by obtaining user feedback.

[0065] An analysis unit uses a transfer learning strategy to perform requirement analysis on the usage data and requirement feedback information.

[0066] A game unit conducts a game between the result of value analysis and the occupation of resources to generate a final software requirement prediction result.

[0067] If the above functions are implemented in the form of software function units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs, etc., all kinds of media that can store program codes.

[0068] The logic and / or steps represented in the flowchart or described in other ways herein, for example, can be considered as a defined sequence list of executable instructions for implementing logical functions, and can be specifically implemented in any computer-readable medium for use by an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processor, or other systems that can fetch and execute instructions from the instruction execution system, apparatus, or device), or in combination with these instruction execution systems, apparatus, or devices. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transmit a program for use by or in combination with an instruction execution system, apparatus, or device.

[0069] More specific examples (non-exhaustive list) of computer-readable media include the following: electrical connection parts with one or more wirings (electronic devices), portable computer disk cartridges (magnetic devices), random access memories (RAM), read-only memories (ROM), erasable programmable read-only memories (EPROM or flash memories), fiber optic devices, and portable compact disc read-only memories (CDROM). Additionally, the computer-readable medium can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other media, then editing, interpreting, or processing it in other suitable ways as necessary, and then storing it in a computer memory.

[0070] It should be understood that each part of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, any one or a combination of the following techniques well known in the art can be used: discrete logic circuits having logic gate circuits for implementing logical functions on data signals, application specific integrated circuits having appropriate combinational logic gate circuits, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), etc.

[0071] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered by the scope of the claims of the present invention.

Claims

1. An AI-based intelligent analysis and prediction method for software requirements, characterized in that, include: When the software product is in use, the software usage data is collected; and demand feedback information is generated by obtaining user feedback; Using the transfer learning strategy, the usage data is analyzed for demand; based on the demand feedback information, the software functions obtained from the demand analysis are matched twice, and based on the results of the secondary matching, each end point output by the deep learning algorithm is analyzed for value; at the same time, based on the usage data, the combined value analysis of the existing functions of the software is directly performed; Compare the results of value analysis with resource usage to generate the final software demand forecast results: select the alternative function combinations as the forecast results, select high-quality combinations and mark them, and generate an analysis report; The analysis report records each prediction result, the value analysis result of the prediction result and the resources occupied; The transfer learning strategy includes obtaining different software products, setting : The i-th software product is ; Starting from the main menu as the starting node, taking each interactive sub-menu as a node, and taking each software function as the end point, the connection lines represent the interaction relationships; respectively construct the paths of this software and for each function implementation as the function map; When a user exits an interactive behavior at a node other than the end point or the start point, it is recorded as an implicit event. Statistical frequency of occurrence of recessive events: ; where represents the frequency of occurrence of recessive events at node b; represents the number of occurrences of recessive events at node b; represents the total number of times reaching node b through interaction; Predict the requirements of this software through a pre-trained deep learning algorithm Input: and Functional maps and the frequency of each latent event occurrence; Output: Endpoints that are in but not in, and at the same time output the probability that each endpoint represents a software requirement, to obtain the endpoint set Z The secondary matching includes, after extracting keywords from the demand feedback information, calculating the distance with each endpoint in the endpoint set Z; for each endpoint, multiplying the similarity by the number of each keyword, and then summing the similarities to obtain the value of each endpoint in the endpoint set Z, where each endpoint represents a demand selection; the similarity is the degree of association between the demand selection represented by each endpoint and each keyword; The combined value analysis includes assuming that the user simultaneously uses m functions, expressed as ; randomly combining the m software and calculating the value of each function combination: ; where represents the value of the y-th combination; represents the probability that the functions represented by the n-th element in the combination are used while the functions represented by the elements in the y-th combination are used simultaneously; n represents the number of elements in the y-th combination; represents the expansion coefficient, ; each function combination represents a function selection; The occupation of the said resources includes that for each function selection, on the basis of the existing functions in this software realized, resources need to be occupied for calculation; Through a pre-trained machine learning algorithm, calculate the resource occupancy for each function selection; Input: function selection and the existing functions; Output: the resources required to be occupied by the function selection.

2. The AI-based intelligent analysis and prediction method for software requirements according to claim 1, wherein: The usage data includes the user's operation records in the software, the query records of the search box, and the records of the user's simultaneous use of multiple software functions of the software; Each record includes the time when the record occurred.

3. The AI-based intelligent software requirement analysis and prediction method according to claim 2, characterized in that: The demand feedback information includes text information on software requirements provided by users through feedback channels.

4. The AI-based intelligent software requirement analysis and prediction method according to claim 3, wherein: The game between the result of value analysis and the occupation of resources includes obtaining all function options, which are set to G; randomly extracting the G function options, with the maximum number of each extraction being G, and the combination of function options extracted each time being different; Get Q function selection combinations, and calculate the total value and total occupied resources of each function selection combination; After eliminating the function selection combinations whose total occupied resources exceed λ, an alternative function selection combination is obtained; wherein λ represents a preset total occupied resource threshold; The high-quality combination includes, for each prediction result, playing a game with the goal of maximizing the total value and minimizing the total occupied resources; establishing a two-dimensional coordinate system with the horizontal axis representing the total occupied resources and the vertical axis representing the total value; Multiply the total value of each prediction result by and use it as the ordinate; the total occupied resources of each prediction result are used as the abscissa; deploy each prediction result in a two-dimensional coordinate system; Among them, = the maximum value of the total value in the prediction result / , so that the maximum values of the horizontal and vertical coordinates are both , if the prediction result falls within , , the triangle formed by the three points, it is judged as an excellent combination.

5. An AI-based software requirements intelligent analysis and prediction system using the method according to any one of claims 1 to 4, characterized in that: The collection unit collects the usage data of the software when the software product is used; and generates demand feedback information by obtaining user feedback; An analysis unit, using a transfer learning strategy to perform demand analysis on the usage data and demand feedback information; The game unit plays a game between the result of value analysis and the occupation of resources to generate the final software demand forecast result.

6. A computer device, comprising: Memory and processor; The memory stores a computer program, characterized in that: when the processor executes the computer program, the steps of the method according to any one of claims 1-4 are implemented.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, the steps of the method according to any one of claims 1-4 are implemented.

Images