Business test information generation method and device and computer program product

Abstract

The invention discloses a service test information generation method and device and a computer program product. The method comprises the steps of determining a target test dimension of service test information, and performing feature dimension reduction on questions in a preset question bank based on the target test dimension to obtain a target question bank of the target test dimension; questions are screened from the target question bank through a hybrid search algorithm, and a candidate question pool is generated; an expected test index of the service test information is determined, questions in the candidate question pool are adjusted through the expected test index, an updated candidate question pool is obtained, and the expected test index comprises at least one of a question difficulty level proportion, a question type proportion and a question score proportion; and determining a target question pool from the updated candidate question pools, and generating target service test information based on the target question pool. Through the method and the device, the problem of low generation efficiency of the service test information in related technologies is solved.

Description

Technical Field

[0001] The present application relates to the field of business testing, and in particular to a method, device, and computer program product for generating business testing information. Background Art

[0002] Employees undergoing business training are required to take a business test at the end of the training to determine the effectiveness of their training. Different test questions are used for each employee. However, since different types of business test questions vary in difficulty, type, and score, ensuring consistency in the proportions of each dimension when randomly selecting questions becomes a pressing challenge. Related technologies manually combine various business test questions to generate test papers, but this method of generating test papers is inefficient.

[0003] Currently, no effective solution has been proposed to the problem of low efficiency in generating business test information in related technologies. Summary of the Invention

[0004] The main purpose of this application is to provide a method, device and computer program product for generating business test information to solve the problem of low efficiency in generating business test information in related technologies.

[0005] To achieve the above-mentioned objectives, according to one aspect of the present application, a method for generating business test information is provided. The method comprises: determining a target test dimension for the business test information, performing feature dimensionality reduction on questions in a preset question bank based on the target test dimension to obtain a target question bank for the target test dimension; screening questions from the target question bank using a hybrid search algorithm to generate a candidate question pool; determining expected test indicators for the business test information, adjusting questions in the candidate question pool based on the expected test indicators to obtain an updated candidate question pool, wherein the expected test indicators include at least one of the following: question difficulty level ratio, question type ratio, and question score ratio; determining a target question pool from the updated candidate question pool, and generating target business test information based on the target question pool.

[0006] Optionally, screening questions from a target question bank through a hybrid search algorithm to generate a candidate question pool includes: determining an initial population through a genetic algorithm, wherein the initial population includes multiple individuals, and each individual represents a group of question combinations; screening target individuals from the initial population, wherein the similarity between the test indicators of the target individuals and the expected test indicators is greater than or equal to a similarity threshold; constructing a search tree with the target individuals as the starting point through a Monte Carlo tree algorithm, and generating multiple groups of target question combinations based on the search tree; using each group of target question combinations as individuals of an updated population, iteratively processing the updated population through a genetic algorithm to obtain multiple updated target individuals; determining a preset fitness function, screening the target individual with the maximum fitness from the updated multiple target individuals through the fitness function, and determining the question combination corresponding to the target individual with the maximum fitness as the candidate question pool.

[0007] Optionally, adjusting the questions in the candidate question pool by the expected test indicators to obtain an updated candidate question pool includes: extracting multiple target indicators from the expected test indicators, and determining the expected indicator value of each target indicator, wherein the target indicator includes at least one of the following: question difficulty level ratio, question type ratio and question score ratio; determining the current indicator value of each target indicator in the candidate question pool, and calculating the deviation of each target indicator based on the current indicator value and the expected indicator value; inputting the deviation of each target indicator into the proportional integral differential control model to obtain a question adjustment strategy; adjusting the questions in the candidate question pool by the question adjustment strategy to obtain an updated candidate question pool.

[0008] Optionally, after generating the candidate question pool, the method further includes: calculating the similarity between each two questions in the candidate question pool by using a preset similarity function; when the similarity is greater than or equal to a similarity threshold, determining the two questions corresponding to the similarity as a group of questions to be optimized; removing any one question in each group of questions to be optimized from the candidate question pool, screening new questions from the question bank according to the expected test indicators, and adding the new questions to the candidate question pool.

[0009] Optionally, determining the target question pool from the updated candidate question pool includes: determining each question in the question bank as a node, and connecting the edges between the nodes based on the mutual exclusion relationship and dependency relationship between the questions, and constructing a preset knowledge graph based on all nodes and all edges; determining the node corresponding to each question in the candidate question pool in the preset knowledge graph, obtaining multiple target nodes, and judging whether there are edges between the multiple target nodes; in the case where there are edges between multiple target nodes, for the edges representing the mutual exclusion relationship, deleting the questions corresponding to any one of the two nodes connected by the edge from the candidate question pool, screening new questions from the question bank according to the expected test indicators, and adding the new questions to the candidate question pool; for the edges representing the dependency relationship, adjusting the order of the questions corresponding to the two nodes connected by the edge based on the dependency relationship; in the case where there is no edge between multiple target nodes, and there is no edge representing the dependency relationship between the target node and other nodes, the candidate question pool is determined as the target question pool.

[0010] Optionally, when there are edges between multiple target nodes, after adjusting the order of questions corresponding to two nodes connected by the edges based on the dependency relationship, the method also includes: for target nodes to which no edges are connected, determining whether there are other nodes connected to the target node in the preset knowledge graph, wherein the other nodes are nodes other than the multiple target nodes in the preset knowledge graph; when there are other nodes connected to the target node in the preset knowledge graph, and the edges between the target node and the other nodes represent a dependency relationship, adjusting the questions corresponding to the target node according to the expected test indicators.

[0011] Optionally, after determining whether there are edges between multiple target nodes, the method also includes: in the case that there are no edges between multiple target nodes, determining whether there are other nodes connected to the target node in the preset knowledge graph, wherein the other nodes are nodes other than the multiple target nodes in the preset knowledge graph; in the case that there are other nodes connected to the target node in the preset knowledge graph, and the edges between the target node and the other nodes represent dependency relationships, adjusting the questions corresponding to the target node according to the expected test indicators; in the case that there are no edges representing dependency relationships between the target node and the other nodes, determining the candidate question pool as the target question pool.

[0012] To achieve the above-mentioned purpose, according to another aspect of the present application, a device for generating business test information is provided. The device includes: a first determination unit, configured to determine a target test dimension of the business test information, and perform feature dimensionality reduction on questions in a preset question bank based on the target test dimension to obtain a target question bank of the target test dimension; a screening unit, configured to screen questions from the target question bank using a hybrid search algorithm to generate a candidate question pool; a second determination unit, configured to determine expected test indicators of the business test information, and adjust the questions in the candidate question pool using the expected test indicators to obtain an updated candidate question pool, wherein the expected test indicators include at least one of the following: question difficulty level ratio, question type ratio, and question score ratio; and a third determination unit, configured to determine a target question pool from the updated candidate question pool, and generate target business test information based on the target question pool.

[0013] In order to achieve the above-mentioned purpose, according to another aspect of the present application, a computer program product is provided, including a computer program, which, when executed by a processor, implements the steps of the method for generating service test information described in each embodiment of the present application.

[0014] Through this application, the following steps are adopted: determining the target test dimension of business test information, performing feature dimensionality reduction on the questions in the preset question bank based on the target test dimension, and obtaining the target question bank of the target test dimension; screening questions from the target question bank through a hybrid search algorithm to generate a candidate question pool; determining the expected test index of business test information, adjusting the questions in the candidate question pool through the expected test index to obtain an updated candidate question pool, wherein the expected test index includes at least one of the following: question difficulty level ratio, question type ratio and question score ratio; determining the target question pool from the updated candidate question pool, and generating target business test information based on the target question pool, thereby solving the problem of low efficiency in generating business test information in related technologies. By reducing the dimensionality of the questions in the question bank based on the target test dimension, it is convenient to adjust the question ratio of each dimension, dynamically screening questions through a hybrid search algorithm, improving the generation efficiency of the question pool, adjusting the questions based on the expected test index, ensuring the consistency of each generated target business test information in dimensions such as difficulty, type and score, and thus achieving the effect of improving the generation efficiency of business test information. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The accompanying drawings, which constitute part of this application, are intended to provide a further understanding of this application. The exemplary embodiments and descriptions of this application are intended to explain this application and do not constitute an improper limitation on this application. In the accompanying drawings:

[0016] Figure 1 This is a flowchart of a method for generating service test information according to an embodiment of the present application;

[0017] Figure 2 Schematic diagram of a test paper generation method provided according to an embodiment of the present application;

[0018] Figure 3 is a schematic diagram of a feature matching process provided according to an embodiment of the present application;

[0019] Figure 4 is a schematic diagram of a device for generating service test information according to an embodiment of the present application;

[0020] Figure 5 is a schematic diagram of an electronic device provided according to an embodiment of the present application. DETAILED DESCRIPTION

[0021] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in this application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0022] In order to enable those skilled in the art to better understand the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of this application.

[0023] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present application described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0024] The present invention will be described below in conjunction with preferred implementation steps. Figure 1 This is a flow chart of a method for generating service test information according to an embodiment of the present application. Figure 1 As shown, the method includes the following steps:

[0025] Step S101 : determining a target test dimension of the business test information, performing feature dimensionality reduction on questions in a preset question bank based on the target test dimension, and obtaining a target question bank of the target test dimension.

[0026] In step S101, the business test information may be a test paper used to assess employees undergoing business training. The target test dimensions may include knowledge points, difficulty, question type, cognitive level, etc. All questions in the question bank are classified according to difficulty level (e.g., easy, medium, difficult), type (e.g., multiple-choice questions, fill-in-the-blank questions, essay questions), and score (e.g., 1 point, 2 points, 3 points), and a unique identifier is assigned to each question. Based on business test requirements, weight values are set for each target test dimension, such as difficulty weight, type weight, and score weight.

[0027] Tensor decomposition is used to reduce the high-dimensional features of the questions in the question bank. During the generation of the target question bank, constraints (such as knowledge points, difficulty, question type, and cognitive level) are set to form a high-dimensional tensor. Core features are extracted through CP (CANDECOMP / PARAFAC decomposition, a nonlinear tensor decomposition method) / Tucker (a linear tensor decomposition method) to reduce computational complexity. Question attributes are constructed as a multidimensional tensor (such as "knowledge point × difficulty × question type").

[0028] Step S102: Screen questions from the target question bank using a hybrid search algorithm to generate a candidate question pool.

[0029] In step S102, the hybrid search algorithm can be a hybrid algorithm of MCTS (Monte Carlo Tree Search) and genetic algorithm, which prepares a structured question bank based on preset target test dimensions, such as difficulty, type, knowledge points, etc. The total number of questions in the test paper, the target proportion of each test dimension, and any additional constraints (such as mutual exclusion rules between questions) are clarified. A population containing different question combinations is initialized using a genetic algorithm, with each population member representing a potential test paper. Ensure that the question combinations in the initial population cover a wide area of the target question bank to increase the diversity of the search.

[0030] Based on the distribution of the test papers along the target test dimension, the best-performing individuals are selected from the genetic algorithm population. Using these high-quality individuals as a starting point, the MCTS search is conducted to deeply explore possible question combination paths. The distribution of questions within the test papers is strategically evaluated to ensure compliance with the test objectives. During the MCTS simulation phase, well-performing test papers are cross-pollinated to generate new test papers to explore more possible question combinations. Certain questions within the test papers are randomly altered to break local optima and increase search flexibility and innovation. MCTS assigns a score to each question combination based on the simulation results, and these scores serve as reward values.

[0031] The MCTS reward value is fed back to the genetic algorithm to update the fitness assessment of the population members, guide the evolution of the genetic algorithm, and prioritize the retention and optimization of those problem combinations with better performance in MCTS. The genetic algorithm performs crossover and mutation operations based on the updated fitness function, and then selects the best-performing individuals to enter the next generation of the population. High-quality individuals are selected from the updated population and the MCTS deep search and optimization are performed again, forming a positive feedback loop that continuously enhances the quality and diversity of solutions. After multiple rounds of iterative optimization, the problem combinations that perform best on the target test dimension are selected from the genetic algorithm population. The problems in these combinations are merged to form a candidate problem pool. Ensure that the distribution of problems in the pool across each dimension meets the test objectives and remove duplicate problems.

[0032] Step S103, determining the expected test indicators of the business test information, adjusting the questions in the candidate question pool according to the expected test indicators, and obtaining an updated candidate question pool, wherein the expected test indicators include at least one of the following: question difficulty level ratio, question type ratio, and question score ratio.

[0033] In step S103, the expected test indicators can be the total number of questions required for the test paper and the specific requirements for each dimension, such as the number of questions of each difficulty level, type, and score. The target ratio (question difficulty level ratio, question type ratio, and question score ratio) for each target test dimension is calculated to ensure that the final test paper meets the expected test indicators in each dimension.

[0034] For example, initialize an empty test paper list. Randomly extract questions from the question bank based on the expected test indicators and the hybrid search algorithm, while taking into account the distribution of existing questions in the current test paper to ensure that the proportion requirements of each dimension are met after each extraction. If it is difficult to find a qualified question in a certain dimension, adjust the proportion of other dimensions appropriately, but still ensure the consistency of the overall proportion. When the number of questions in the test paper reaches the predetermined total, stop the question extraction process. Verify the generated test paper to check whether it meets the proportion requirements of each dimension. If not, return to re-examine the questions. Output the final test paper and save the relevant data for subsequent use or analysis.

[0035] Step S104: determining a target topic pool from the updated candidate topic pool, and generating target business test information based on the target topic pool.

[0036] In step S104, the updated candidate question pool undergoes a compliance check to ensure that there are no logical conflicts between questions and that all questions meet the expected test criteria. Once the candidate question pool meets all compliance and metric requirements, it is identified as the target question pool, which serves as the basis for generating specific test information. Based on the pre-set test paper template, the questions in the target question pool are pasted into the test paper to generate the target business test information.

[0037] The method for generating business test information provided by the embodiment of the present application determines the target test dimension of the business test information, performs feature dimensionality reduction on the questions in the preset question bank based on the target test dimension, and obtains the target question bank of the target test dimension; screens the questions from the target question bank through a hybrid search algorithm to generate a candidate question pool; determines the expected test index of the business test information, adjusts the questions in the candidate question pool through the expected test index, and obtains an updated candidate question pool, wherein the expected test index includes at least one of the following: question difficulty level ratio, question type ratio and question score ratio; determines the target question pool from the updated candidate question pool, and generates target business test information based on the target question pool, thereby solving the problem of low efficiency in generating business test information in related technologies. By reducing the dimensionality of the questions in the question bank based on the target test dimension, it is convenient to adjust the question ratio of each dimension, dynamically screens the questions through a hybrid search algorithm, improves the generation efficiency of the question pool, adjusts the questions based on the expected test index, and ensures the consistency of each generated target business test information in dimensions such as difficulty, type and score, thereby achieving the effect of improving the generation efficiency of business test information.

[0038] In order to improve the efficiency of generating business test information, a hybrid search algorithm is used to screen questions to generate a candidate question pool. Optionally, in the method for generating business test information provided in an embodiment of the present application, questions are screened from a target question bank through a hybrid search algorithm, and generating a candidate question pool includes: determining an initial population through a genetic algorithm, wherein the initial population includes multiple individuals, and each individual represents a group of question combinations; screening out target individuals from the initial population, wherein the similarity between the test indicators of the target individuals and the expected test indicators is greater than or equal to a similarity threshold; constructing a search tree with the target individuals as the starting point through a Monte Carlo tree algorithm, and generating multiple groups of target question combinations based on the search tree; using each group of target question combinations as individuals of an updated population, iteratively processing the updated population through a genetic algorithm to obtain multiple updated target individuals; determining a preset fitness function, screening out a target individual with the maximum fitness from the updated multiple target individuals through the fitness function, and determining the question combination corresponding to the target individual with the maximum fitness as the candidate question pool.

[0039] In some embodiments, multiple individuals are randomly generated, each individual is a question combination, and this combination can be formed by randomly selecting questions from the target question bank. The size of the population needs to be large enough to cover the diversity in the question bank. The individual is represented by a vector or matrix, and each element in the vector or matrix represents the characteristics of the question (such as difficulty, type, score) and whether it is selected. Create a similarity function to quantify the similarity between the test indicators of the individual (question combination) and the expected test indicators. Determine the similarity threshold, and only when the similarity between the individual's test indicators and the expected test indicators is greater than or equal to the threshold will it be regarded as a target individual. From the initial population, screen out individuals that meet or exceed the similarity threshold according to the similarity function.

[0040] The selected target individual is selected as the starting node of the Monte Carlo tree algorithm. Through Monte Carlo simulation, the search tree is gradually expanded to explore other problem combinations that can be achieved starting from the current individual, while the performance of each combination is evaluated based on the preset fitness function. In the search tree, multiple groups of the best-performing problem combinations are selected based on fitness. The target problem combinations generated from the Monte Carlo tree algorithm are added to the population as a new generation of individuals. The selection, crossover, and mutation operations in the genetic algorithm are performed to iteratively optimize the population to find problem combinations that better meet the expected test indicators. The evaluation and screening process is repeated, and the target individual is again selected from the updated population.

[0041] From the updated target individuals, the individual with the maximum fitness is selected by comparing the values of the fitness function. The question combination of the target individual with the maximum fitness is determined as the candidate question pool. This set of question combinations is closest to the expected test indicators in all dimensions and has the best fitness in the population.

[0042] Among them, the fitness function can be expressed by the following formula:

[0043] Fitness = 1 / (1+F+Penalty);

[0044] Penalty=η*ΣViolation_Count;

[0045] Where Penalty is the sum of η (the intermediate parameter) and Violation_Count (the number of times the constraint is violated). F is the target optimization function. The higher the fitness, the closer the individual is to the expected test index, while the smaller F and Penalty are, the higher the fitness. In a genetic algorithm, the fitness of an individual (i.e., a test paper solution) is determined by its F value and the number of times the constraint is violated. F is determined by the following formula:

[0046] Minimize F=γ1*|C_real-C_target|+γ2*|D_real-D_target|+γ3*Var(s);

[0047] Subject to:

[0048] Σs_i=TotalScore;

[0049]

[0050] Here, C_real is the actual total score of all questions in the candidate pool, C_target is the target total score, D_real is the actual difficulty of all questions in the candidate pool, D_target is the target difficulty of all questions in the candidate pool, and Var(s) is the variance of the scores, which refers to the difference in scores between questions. γ1 is the score weight, γ2 is the difficulty weight, and γ3 is the type weight. There are two constraints: the total score of all questions in the candidate pool is TotalScore, and Q_i (question) is not in Conflict_Set (question conflict set), meaning that certain questions cannot appear on the same test paper at the same time.

[0051] Through the hybrid search algorithm, questions are intelligently screened from the target question bank. The generated candidate question pool not only meets the expected indicators of business testing, but also achieves a balance between global optimization and local optimization through the mutual cooperation of genetic algorithm and Monte Carlo tree algorithm, making the business test information design both reasonable and efficient.

[0052] In order to ensure that the candidate question pool meets the expected test indicators, a proportional-integral-differential control strategy is used to adjust the questions in the candidate question pool. Optionally, in the method for generating business test information provided in an embodiment of the present application, the questions in the candidate question pool are adjusted according to the expected test indicators to obtain an updated candidate question pool, including: extracting multiple target indicators from the expected test indicators, and determining the expected indicator value of each target indicator, wherein the target indicator includes at least one of the following: a question difficulty level ratio, a question type ratio, and a question score ratio; determining the current indicator value of each target indicator of the candidate question pool, and calculating the deviation of each target indicator based on the current indicator value and the expected indicator value; inputting the deviation of each target indicator into a proportional-integral-differential control model to obtain a question adjustment strategy; adjusting the questions in the candidate question pool according to the question adjustment strategy to obtain an updated candidate question pool.

[0053] In some embodiments, the ratio of difficulty, type, and score of questions in the current candidate question pool is analyzed, and the current indicator value for each dimension is recorded. The current indicator value is compared with the expected indicator value, and the degree of deviation of each target indicator is calculated, that is, the difference between the actual ratio and the preset ratio. According to the test requirements, a proportional-integral-differential control model (proportional coefficient P, integral coefficient I, differential coefficient D) is set to meet the sensitivity and stability requirements of the question pool adjustment. The deviation of each target indicator is used as input, and the proportional-integral-differential control model is run. Through the comprehensive calculation of the proportional term, integral term, and differential term, an adjustment strategy is obtained. The proportional term quickly responds to deviations, the integral term eliminates static errors, and the differential term prevents over-adjustment. Based on the adjustment strategy of the proportional-integral-differential model, questions of specific types, difficulties, or scores are added or deleted from the question pool to reduce deviations. For example, if the difficulty ratio deviation is large, it can be adjusted by adding or removing questions of the corresponding difficulty. For minor deviations, a question replacement strategy can also be adopted, that is, replacing questions with similar questions that better meet the ratio requirements. After executing the adjustment strategy, all modifications are integrated to generate a new question combination, thereby obtaining an updated candidate question pool.

[0054] This embodiment uses a proportional-integral-derivative control model to adjust and optimize the candidate question pool, ensuring that the composition of the question pool more accurately matches the test requirements, thereby improving the effectiveness and fairness of the test. It can process large amounts of data in a short period of time and make efficient and accurate question pool adjustment decisions.

[0055] Similar questions need to be eliminated from the candidate question pool. Optionally, in the method for generating business test information provided in the embodiment of the present application, after generating the candidate question pool, the method further includes: calculating the similarity between each two questions in the candidate question pool by using a preset similarity function; when the similarity is greater than or equal to a similarity threshold, determining the two questions corresponding to the similarity as a group of questions to be optimized; eliminating any one question in each group of questions to be optimized from the candidate question pool, screening new questions from the question bank according to the expected test indicators, and adding the new questions to the candidate question pool.

[0056] In some embodiments, the similarity function can be expressed by the following formula:

[0057] Sim(Q_i, Q_j)=[λ1|d_i-d_j|+λ2||t_i-t_j||_2+λ3*JS(k_i||k_j)];

[0058] Here, λ1+λ2+λ3=1, and the value of Sim should be between 0 and 1. When all three difference terms are large, Sim approaches 0, and vice versa. These three difference terms are: the absolute difference in question difficulty d, the Euclidean distance of the knowledge point distribution t, and the JS divergence between the knowledge point distributions k_i and k_j. For each pair of questions in the candidate question pool, calculate their similarity using the similarity function described above. The results can be stored in a matrix for subsequent analysis and processing.

[0059] A similarity threshold is set. If the calculated similarity is greater than or equal to the threshold, questions Q_i and Q_j are considered a group of questions to be optimized. The two questions are too similar in content, structure, or attributes, which may lead to duplication or redundancy of test information. For each group of questions to be optimized: a question is randomly selected from the group and removed to reduce redundancy between questions. New questions are selected from the question bank based on the expected test indicators. These new questions are added to the candidate question pool to fill the gaps left by the removal of similar questions.

[0060] This embodiment ensures the diversity of the question pool by optimizing similar questions, preventing overly similar questions from appearing, and can maintain the consistency of the question pool with the expected test indicators while eliminating redundancy, thereby improving the overall quality and efficiency of the test.

[0061] In order to avoid conflicts in the candidate question pool, the candidate question pool needs to be updated based on the preset knowledge graph. Optionally, in the method for generating business test information provided in the embodiment of the present application, determining the target question pool from the updated candidate question pool includes: determining each question in the question bank as a node, and connecting the edges between the nodes based on the mutual exclusion relationship and dependency relationship between the questions, and constructing a preset knowledge graph based on all nodes and all edges; determining the corresponding node of each question in the candidate question pool in the preset knowledge graph, obtaining multiple target nodes, and judging whether there are edges between the multiple target nodes; in the case where there are edges between multiple target nodes, for the edges representing the mutual exclusion relationship, deleting the questions corresponding to any one of the two nodes connected by the edge from the candidate question pool, screening new questions from the question bank according to the expected test indicators, and adding the new questions to the candidate question pool; for the edges representing the dependency relationship, adjusting the order of the questions corresponding to the two nodes connected by the edge based on the dependency relationship; in the case where there is no edge between the multiple target nodes, and there is no edge representing the dependency relationship between the target node and other nodes, the candidate question pool is determined as the target question pool.

[0062] In some embodiments, each question in the question bank is defined as a node, and each node contains attributes such as question ID, question content, difficulty level, question type and score. Based on the mutually exclusive relationship between questions (such as "limited to 2 questions for the same knowledge point") and dependency relationship (such as "question A must be before question B"), the edges between the nodes are connected. A mutually exclusive relationship means that two questions cannot appear in the same test at the same time; a dependency relationship means that a question needs to be tested after another question to ensure the logic and consistency of the test process. Determine the corresponding nodes in the preset knowledge graph for each question in the updated candidate question pool, and these nodes are the target nodes. Check whether there are edges between the target nodes that represent mutually exclusive relationships. If it is found that a mutually exclusive edge connects two target nodes, delete the question corresponding to one of the nodes from the candidate question pool to eliminate the mutually exclusive relationship. Filter out new questions from the question bank according to the expected test indicators to fill the gaps caused by the deletion of questions, and ensure the integrity of the question pool and the consistency of the test indicators.

[0063] Check the target node and other nodes in the question pool for edges that represent dependencies. Based on these dependencies, adjust the order of questions in the test to ensure that all questions are tested in the correct logical order, avoiding knowledge jumps or incoherence. After ensuring that there are no mutually exclusive edges between all target nodes and no unprocessed dependency edges between the target node and other nodes, check whether the target node meets the expected test criteria. If all questions in the current candidate question pool meet the expected criteria and the relationships between questions have been properly handled, the current candidate question pool becomes the target question pool.

[0064] By constructing and utilizing a knowledge graph, this embodiment can systematically identify and process the mutual exclusion and dependency relationships between questions, ensuring that the question pool not only meets the test indicators, but also meets expectations in terms of logic and coherence, significantly improving the efficiency and accuracy of test design.

[0065] Adjust the candidate question pool based on dependency relationships. Optionally, in the method for generating business test information provided in an embodiment of the present application, when there are edges between multiple target nodes, after adjusting the order of questions corresponding to two nodes connected by the edges based on the dependency relationships, the method further includes: for target nodes to which no edges are connected, determining whether there are other nodes connected to the target node in the preset knowledge graph, wherein the other nodes are nodes other than the multiple target nodes in the preset knowledge graph; when there are other nodes connected to the target node in the preset knowledge graph, and the edges between the target node and the other nodes represent dependency relationships, adjusting the questions corresponding to the target node according to the expected test indicators.

[0066] In some embodiments, in the constructed knowledge graph, independent target nodes that are not directly connected to other target nodes through edges are identified. A depth-first or breadth-first search is performed on the knowledge graph to check whether the independent target node is connected to other nodes in the graph through edges. These other nodes can be any nodes other than the target node. The key is to identify edges that represent dependency relationships, that is, the logic that one question depends on another question to be tested first. If an edge with a dependency relationship is found between a target node and other nodes, the order of the questions in the question pool is adjusted according to the direction of this edge to ensure that the dependent questions are tested first, and then the dependent questions are tested.

[0067] If a question depends on multiple other questions, a reasonable order must be determined, and additional sorting algorithms, such as topological sorting, may be needed to ensure that all dependencies are met. If, after adjusting the order, it is found that some questions require specific knowledge points or skills as prerequisites, but these knowledge points or skills are not covered in the current question pool, new questions that meet the expected test indicators must be selected from the question bank to fill the gaps in content. If the dependencies between questions cause some questions to no longer be suitable for the current test design, the questions may need to be replaced to ensure that the test content meets both the logical order and the expected test indicators.

[0068] This embodiment adjusts the questions in the candidate question pool based on dependency relationships to ensure that the logical order of the questions in the question pool is correct. While meeting specific test indicators, it maintains the comprehensiveness and professionalism of the test content, providing a guarantee for high-quality test information preparation.

[0069] Optionally, in the method for generating business test information provided in an embodiment of the present application, after determining whether there are edges between multiple target nodes, the method further includes: in the case that there are no edges between multiple target nodes, determining whether there are other nodes connected to the target node in the preset knowledge graph, wherein the other nodes are nodes other than the multiple target nodes in the preset knowledge graph; in the case that there are other nodes connected to the target node in the preset knowledge graph, and the edges between the target node and the other nodes represent dependency relationships, adjusting the questions corresponding to the target node according to the expected test indicators; in the case that there are no edges representing dependency relationships between the target node and the other nodes, determining the candidate question pool as the target question pool.

[0070] In some embodiments, it is confirmed which target nodes are independent in the current candidate topic pool, that is, they are not directly connected to other target nodes through edges. For each independent target node, check whether it is connected to other nodes (non-target nodes) in the knowledge graph through edges. If an edge with a dependency relationship is found between a target node and other non-target nodes, this indicates that the question depends on the prior knowledge or skills of other specific questions in the graph. Based on the dependencies, the order of the questions is adjusted to ensure that all dependencies are met, that is, any dependent questions appear before the questions that depend on them. According to the adjusted order, check whether the topic pool completely covers all dependencies. If some dependent prerequisite knowledge is missing in the topic pool, filter and add corresponding questions from the question bank according to the expected test indicators to fill the content gaps.

[0071] If there are no edges representing dependencies between the target node and other non-target nodes, this means that the current pool of candidate questions is logically independent and requires no additional ordering adjustments. After confirming that all external dependencies of the target nodes have been properly handled, or that no external dependencies exist, the current candidate question pool can be directly determined as the target question pool for test design.

[0072] This embodiment adjusts the questions in the candidate question pool based on dependency relationships, ensuring the logical coherence of the question pool and the satisfaction of test indicators even in the absence of direct connections between target nodes, thereby providing a comprehensive and professional foundation for test design.

[0073] According to another embodiment of the present application, a test paper generation method is also provided. Figure 2 Schematic diagram of the test paper generation method provided in accordance with the embodiment of the present application. Figure 2 As shown, the method includes: question feature extraction, competition requirement input, initial question screening, feature matching, dynamic adjustment algorithm, test paper generation and test paper evaluation.

[0074] in, Figure 3 This is a schematic diagram of the feature matching process provided in accordance with an embodiment of the present application. Figure 3 As shown in the figure, the process includes: multi-source question bank management, dynamic feature matching, genetic optimization processing and test paper quality verification.

[0075] This example builds an efficient and adaptive question generation system through multi-dimensional constraint dimensionality reduction, dynamic strategy control, hybrid search efficiency enhancement, and knowledge-driven conflict resolution. It dynamically adjusts assessment content based on the user's ability level and learning behavior, enabling precise and personalized learning effect evaluation and providing real-time feedback and optimization suggestions.

[0076] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0077] The present application also provides a device for generating service test information. It should be noted that the device for generating service test information in the present application can be used to execute the method for generating service test information provided in the present application. The device for generating service test information provided in the present application is described below.

[0078] Figure 4 Schematic diagram of a device for generating service test information according to an embodiment of the present application. Figure 4 As shown, the device includes:

[0079] The first determining unit 401 is configured to determine a target test dimension of the business test information, and perform feature dimensionality reduction on questions in a preset question bank based on the target test dimension to obtain a target question bank of the target test dimension;

[0080] A screening unit 402 is configured to screen questions from a target question bank using a hybrid search algorithm to generate a candidate question pool;

[0081] A second determining unit 403 is configured to determine an expected test indicator of the business test information, and adjust the questions in the candidate question pool according to the expected test indicator to obtain an updated candidate question pool, wherein the expected test indicator includes at least one of the following: a question difficulty level ratio, a question type ratio, and a question score ratio;

[0082] The third determining unit 404 is configured to determine a target topic pool from the updated candidate topic pool, and generate target business test information based on the target topic pool.

[0083] The device for generating business test information provided in an embodiment of the present application includes a first determining unit 401 that determines a target test dimension of the business test information, performs feature dimensionality reduction on questions in a preset question bank based on the target test dimension, and obtains a target question bank of the target test dimension. A screening unit 402 screens questions from the target question bank using a hybrid search algorithm to generate a candidate question pool. A second determining unit 403 determines expected test indicators for the business test information, adjusts questions in the candidate question pool based on the expected test indicators, and obtains an updated candidate question pool, wherein the expected test indicators include at least one of the following: question difficulty level ratio, question type ratio, and question score ratio. A third determining unit 404 determines a target question pool from the updated candidate question pool and generates target business test information based on the target question pool. This solves the problem of low business test information generation efficiency in related technologies. By performing dimensionality reduction on questions in the question bank based on the target test dimension, it is convenient to adjust the question ratio of each dimension. Questions are dynamically screened using a hybrid search algorithm to improve the generation efficiency of the question pool. Questions are adjusted based on the expected test indicators to ensure consistency of each generated target business test information in terms of difficulty, type, and score, thereby achieving the effect of improving the generation efficiency of business test information.

[0084] Optionally, in the device for generating business test information provided in an embodiment of the present application, the screening unit 402 includes: a first determination module, used to determine an initial population through a genetic algorithm, wherein the initial population includes multiple individuals, and each individual represents a group of question combinations; a first screening module, used to screen out target individuals from the initial population, wherein the similarity between the test indicators of the target individuals and the expected test indicators is greater than or equal to a similarity threshold; a first generation module, used to construct a search tree with the target individuals as the starting point through a Monte Carlo tree algorithm, and generate multiple groups of target question combinations based on the search tree; an iteration module, used to use each group of target question combinations as individuals of an updated population, and iteratively process the updated population through a genetic algorithm to obtain multiple updated target individuals; a second determination module, used to determine a preset fitness function, and screen out the target individual with the maximum fitness from the updated multiple target individuals through the fitness function, and determine the question combination corresponding to the target individual with the maximum fitness as a candidate question pool.

[0085] Optionally, in the device for generating business test information provided in an embodiment of the present application, the second determination unit 403 includes: an extraction module, used to extract multiple target indicators from the expected test indicators, and determine the expected indicator value of each target indicator, wherein the target indicator includes at least one of the following: question difficulty level ratio, question type ratio and question score ratio; a third determination module, used to determine the current indicator value of each target indicator in the candidate question pool, and calculate the deviation of each target indicator based on the current indicator value and the expected indicator value; an input module, used to input the deviation of each target indicator into a proportional integral differential control model to obtain a question adjustment strategy; a first adjustment module, used to adjust the questions in the candidate question pool through the question adjustment strategy to obtain an updated candidate question pool.

[0086] Optionally, in the device for generating business test information provided in the embodiment of the present application, the device also includes: a calculation unit, used to calculate the similarity between each two questions in the candidate question pool through a preset similarity function; a fourth determination unit, used to determine the two questions corresponding to the similarity as a group of questions to be optimized when the similarity is greater than or equal to a similarity threshold; and an elimination unit, used to eliminate any one question in each group of questions to be optimized from the candidate question pool, screen new questions from the question bank according to the expected test indicators, and add the new questions to the candidate question pool.

[0087] Optionally, in the business test information generation device provided in the embodiment of the present application, the third determination unit 404 includes: a fourth determination module, which is used to determine each question in the question bank as a node, and connect the edges between the nodes based on the mutual exclusion relationship and dependency relationship between the questions, and construct a preset knowledge graph based on all nodes and all edges; a fifth determination module, which is used to determine the corresponding node of each question in the candidate question pool in the preset knowledge graph, obtain multiple target nodes, and determine whether there are edges between the multiple target nodes; a first adding module, which is used to, when there are edges between multiple target nodes, for an edge representing a mutually exclusive relationship, delete the question corresponding to any one of the two nodes connected by the edge from the candidate question pool, screen new questions from the question bank according to the expected test indicators, and add the new questions to the candidate question pool; a second adjustment module, which is used to, for an edge representing a dependency relationship, adjust the order of the questions corresponding to the two nodes connected by the edge based on the dependency relationship; a sixth determination module, which is used to determine the candidate question pool as the target question pool when there is no edge between the multiple target nodes and there is no edge representing a dependency relationship between the target node and other nodes.

[0088] Optionally, in the device for generating business test information provided in an embodiment of the present application, the third determination unit 404 also includes: a first judgment module, for judging, for a target node to which no edges are connected, whether there are other nodes connected to the target node in the preset knowledge graph, wherein the other nodes are nodes other than the multiple target nodes in the preset knowledge graph; a third adjustment module, for adjusting the question corresponding to the target node according to the expected test indicator when there are other nodes connected to the target node in the preset knowledge graph and the edges between the target node and the other nodes represent a dependency relationship.

[0089] Optionally, in the device for generating business test information provided in an embodiment of the present application, the third determination unit 404 also includes: a second judgment module, for judging whether there are other nodes connected to the target node in the preset knowledge graph when there are no edges between multiple target nodes, wherein the other nodes are nodes other than the multiple target nodes in the preset knowledge graph; a fourth adjustment module, for adjusting the questions corresponding to the target node according to the expected test indicators when there are other nodes connected to the target node in the preset knowledge graph and the edges between the target node and the other nodes represent a dependency relationship; a seventh determination module, for determining the candidate question pool as the target question pool when there is no edge representing a dependency relationship between the target node and the other nodes.

[0090] The device for generating business test information includes a processor and a memory. The above-mentioned first determination unit 401, screening unit 402, second determination unit 403 and third determination unit 404 are all stored in the memory as program units, and the processor executes the above-mentioned program units stored in the memory to realize corresponding functions.

[0091] The processor contains a kernel, which retrieves the corresponding program unit from the memory. One or more kernels can be set, and the efficiency of generating business test information can be improved by adjusting the kernel parameters.

[0092] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0093] An embodiment of the present invention provides a computer-readable storage medium storing a program, which implements a method for generating service test information when executed by a processor.

[0094] An embodiment of the present invention provides a processor, which is used to run a program, wherein the method for generating business test information is executed when the program is running.

[0095] Figure 5Schematic diagram of an electronic device according to an embodiment of the present application. Figure 5 As shown, the electronic device 501 includes a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, the following steps are implemented: determining the target test dimension of the business test information, performing feature dimensionality reduction on the questions in the preset question bank based on the target test dimension, and obtaining the target question bank of the target test dimension; screening questions from the target question bank using a hybrid search algorithm to generate a candidate question pool; determining the expected test indicators of the business test information, adjusting the questions in the candidate question pool based on the expected test indicators, and obtaining an updated candidate question pool, wherein the expected test indicators include at least one of the following: question difficulty level ratio, question type ratio, and question score ratio; determining the target question pool from the updated candidate question pool, and generating target business test information based on the target question pool. The device herein can be a server, a PC, a PAD, a mobile phone, etc.

[0096] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing a program initialized with the following method steps: determining a target test dimension of business test information, performing feature dimensionality reduction on questions in a preset question bank based on the target test dimension, and obtaining a target question bank of the target test dimension; screening questions from the target question bank through a hybrid search algorithm to generate a candidate question pool; determining expected test indicators of business test information, adjusting questions in the candidate question pool according to the expected test indicators, and obtaining an updated candidate question pool, wherein the expected test indicators include at least one of the following: question difficulty level ratio, question type ratio, and question score ratio; determining a target question pool from the updated candidate question pool, and generating target business test information based on the target question pool.

[0097] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0098] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the steps in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0099] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0100] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0101] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0102] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0103] Computer-readable media includes permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. The information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media (transitory media), such as modulated data signals and carrier waves.

[0104] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0105] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0106] The above are merely embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should all be included within the scope of the claims of the present application.

Claims

1. A method for generating business test information, characterized in that: include: Determine a target test dimension of the business test information, and perform feature dimensionality reduction on questions in a preset question bank based on the target test dimension to obtain a target question bank of the target test dimension; Screening questions from the target question bank using a hybrid search algorithm to generate a candidate question pool; Determining expected test indicators of the business test information, and adjusting the questions in the candidate question pool according to the expected test indicators to obtain an updated candidate question pool, wherein the expected test indicators include at least one of the following: question difficulty level ratio, question type ratio, and question score ratio; A target topic pool is determined from the updated candidate topic pool, and target business test information is generated based on the target topic pool.

2. The method according to claim 1, characterized in that Screening questions from the target question bank using a hybrid search algorithm to generate a candidate question pool includes: Determining an initial population by a genetic algorithm, wherein the initial population includes a plurality of individuals, each individual representing a set of question combinations; Screening a target individual from the initial population, wherein the similarity between the test index of the target individual and the expected test index is greater than or equal to a similarity threshold; Constructing a search tree using the target individual as a starting point through a Monte Carlo tree algorithm, and generating multiple sets of target question combinations based on the search tree; Taking each target question combination as an individual of the updated population, the updated population is iteratively processed by the genetic algorithm to obtain multiple updated target individuals; A preset fitness function is determined, a target individual with the maximum fitness is screened out from the updated multiple target individuals using the fitness function, and a topic combination corresponding to the target individual with the maximum fitness is determined as the candidate topic pool.

3. The method according to claim 1, characterized in that Adjusting the topics in the candidate topic pool according to the expected test indicators to obtain an updated candidate topic pool includes: Extracting a plurality of target indicators from the expected test indicators and determining an expected indicator value for each of the target indicators, wherein the target indicators include at least one of the following: a question difficulty level ratio, a question type ratio, and a question score ratio; Determining a current indicator value of each target indicator in the candidate question pool, and calculating a deviation of each target indicator based on the current indicator value and the expected indicator value; Input the deviation of each target indicator into the proportional integral differential control model to obtain the problem adjustment strategy; The topics in the candidate topic pool are adjusted using the topic adjustment strategy to obtain an updated candidate topic pool.

4. The method according to claim 1, wherein After generating the candidate topic pool, the method further includes: Calculate the similarity between every two topics in the candidate topic pool using a preset similarity function; When the similarity is greater than or equal to a similarity threshold, the two questions corresponding to the similarity are determined as a question group to be optimized; Eliminate any one question in each group of questions to be optimized from the candidate question pool, select new questions from the question bank according to expected test indicators, and add the new questions to the candidate question pool.

5. The method according to claim 1, wherein Determining a target topic pool from the updated candidate topic pool includes: Determine each question in the question bank as a node, connect edges between the nodes based on the mutual exclusion relationship and dependency relationship between the questions, and construct a preset knowledge graph based on all nodes and all edges; Determine a corresponding node in the preset knowledge graph for each question in the candidate question pool, obtain multiple target nodes, and determine whether there are edges between the multiple target nodes; In the case where the edge exists between the multiple target nodes, for the edge representing a mutually exclusive relationship, deleting the question corresponding to any one of the two nodes connected by the edge from the candidate question pool, screening new questions from the question bank according to the expected test indicators, and adding the new questions to the candidate question pool; For an edge representing a dependency relationship, adjusting the order of questions corresponding to two nodes connected by the edge based on the dependency relationship; In a case where the edges do not exist between the multiple target nodes and no edges representing dependency relationships exist between the target node and the other nodes, the candidate topic pool is determined as the target topic pool.

6. The method according to claim 5, characterized in that In the case where the edge exists between the multiple target nodes, after adjusting the order of the questions corresponding to the two nodes connected by the edge based on the dependency relationship, the method further includes: For a target node with no connected edges, determine whether there are other nodes connected to the target node in the preset knowledge graph, wherein the other nodes are nodes in the preset knowledge graph other than the multiple target nodes; When there are other nodes connected to the target node in the preset knowledge graph, and the edges between the target node and the other nodes represent dependency relationships, the questions corresponding to the target node are adjusted according to the expected test indicators.

7. The method according to claim 5, characterized in that After determining whether there is an edge between the multiple target nodes, the method further includes: In the case where the edge does not exist between the multiple target nodes, determining whether there are other nodes connected to the target nodes in the preset knowledge graph, wherein the other nodes are nodes in the preset knowledge graph other than the multiple target nodes; If there are other nodes connected to the target node in the preset knowledge graph, and the edges between the target node and the other nodes represent dependency relationships, adjust the question corresponding to the target node according to the expected test indicators; In the case that there is no edge representing a dependency relationship between the target node and the other nodes, the candidate topic pool is determined as the target topic pool.

8. A device for generating business test information, characterized in that: include: A first determining unit is configured to determine a target test dimension of the business test information, and perform feature dimensionality reduction on questions in a preset question bank based on the target test dimension to obtain a target question bank of the target test dimension; A screening unit, configured to screen questions from the target question bank using a hybrid search algorithm to generate a candidate question pool; a second determining unit, configured to determine an expected test indicator of the business test information, and adjust the questions in the candidate question pool according to the expected test indicator to obtain an updated candidate question pool, wherein the expected test indicator includes at least one of the following: a question difficulty level ratio, a question type ratio, and a question score ratio; The third determining unit is configured to determine a target topic pool from the updated candidate topic pool, and generate target business test information based on the target topic pool.

9. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the method for generating service test information according to any one of claims 1 to 7 is implemented.

10. An electronic device, characterized in that: It includes one or more processors and a memory, the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method for generating business test information as described in any one of claims 1 to 7.

Images