Intelligent virtual simulation method for promoting mental health of women suffering from depression in perinatal period
Through intelligent virtual simulation methods, a personalized virtual game world and social network are built, and the game environment and task design are dynamically adjusted, which solves the problems of low social connection quality and unreasonable task design in the existing technology, and achieves efficient emotional recovery and sense of social participation.
Patent Information
- Application Number
- CN202411655113.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-19
- Publication Date
- 2025-03-07
AI Technical Summary
The existing mental health promotion methods for perinatal depression women have low quality and effectiveness of social connections, and cannot provide personalized social support, resulting in low emotional recovery and social participation, and unreasonable task design leads to the loss of user interest and self-efficacy.
Using intelligent virtual simulation methods, we use the method of collecting user needs, building a personalized virtual game world, dynamically adjusting the game environment, integrating social interaction functions, evaluating task participation and difficulty, analyzing user feedback and interactive data, providing professional psychological counseling, and continuously updating software functions and content.
It improves the quality and effectiveness of social connections, provides personalized social support, enhances the user's emotional recovery speed and sense of social participation, avoids frustration or boredom caused by excessive or excessive tasks, and stimulates users' continuous participation and self-efficacy.
Smart Images
Figure CN119580953A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of psychological medicine, and in particular to an intelligent virtual simulation method for promoting the mental health of women with perinatal depression. Background Art
[0002] Perinatal depression refers to depression that occurs during pregnancy or within one year after childbirth and is a common mental health problem. Perinatal depression (PND) has become an important health problem among women worldwide. Globally, the incidence of perinatal depression is high, seriously affecting women's mental health and quality of life. Depression not only affects women's emotions, cognition, behavior, and social functions, but may also have a negative impact on the health of the fetus and infant. Studies have shown that perinatal depression is closely related to adverse pregnancy outcomes (such as premature birth, low birth weight) and infant developmental delay. Therefore, early identification and intervention of perinatal depression is crucial to improving the health of women and children; therefore, it is particularly important to invent an intelligent virtual simulation method to promote the mental health of women with perinatal depression.
[0003] The quality and effectiveness of social connections in existing methods for promoting mental health in perinatal women with depression are low, and they cannot provide more and more helpful social support to each user, which is not conducive to the emotional recovery and social participation of perinatal women. In addition, existing methods for promoting mental health in perinatal women with depression are prone to frustration or boredom caused by tasks that are too difficult or too easy, which reduces the self-efficacy of perinatal women and the effectiveness of virtual simulation intervention. Therefore, we propose an intelligent virtual simulation method for promoting mental health in perinatal women with depression. Summary of the invention
[0004] The purpose of the present invention is to solve the defects in the prior art and to propose an intelligent virtual simulation method for promoting the mental health of perinatal depressed women.
[0005] In order to achieve the above object, the present invention adopts the following technical solutions:
[0006] An intelligent virtual simulation method for promoting mental health of women with perinatal depression, the specific steps of the method are as follows:
[0007] Ⅰ. Collect the needs of women with perinatal depression and establish the connection between the needs and the software;
[0008] Ⅱ. Build a personalized virtual game world based on user needs;
[0009] Ⅲ. Analyze and manage user emotion changes and dynamically adjust the game environment;
[0010] IV. Integrate social interaction functions to facilitate user communication and optimize the social network structure in real time;
[0011] V. Evaluate the engagement and difficulty of each task and design diverse tasks for each user;
[0012] VI. Analyze user feedback and interaction data and provide professional psychological counseling;
[0013] Ⅶ. Collect user feedback and usage data to evaluate software effectiveness and user satisfaction;
[0014] VIII. Continuously update software functions and content based on user feedback and latest research.
[0015] As a further solution of the present invention, the specific steps of establishing the association between the demand and the software in step I are as follows:
[0016] S1.1: Design a questionnaire covering mental health, social needs, emotional management, and professional support, distribute the questionnaire online or in person, collect the completed questionnaires, conduct statistical analysis on the collected questionnaire data, and identify the main mental health needs and concerns of users;
[0017] S1.2: Conduct in-depth interviews with multiple groups of questionnaire participants through random sampling, record the participants' specific experiences and suggestions, organize the questionnaire and interview data into structured information, obtain the main needs and pain points of users, and encode and classify the interview content;
[0018] S1.3: Extract entities from the sorted demand data through natural language processing technology, including user demand keywords and function module names, and determine the relationship between different entities based on the analysis results. Then use the Neo4j graph database to construct a relationship graph with the extracted entities as nodes and the corresponding relationships between nodes as edges;
[0019] S1.4: Analyze the association graph through graph convolutional networks to explore the potential relationships between groups of nodes, then visualize the constructed association graph through Graphviz, analyze the structure of the graph, identify the relationship between core requirements and software modules, and regularly update the association graph based on new requirement information and changes in software modules.
[0020] As a further solution of the present invention, the specific steps of dynamically adjusting the game environment in step III are as follows:
[0021] S2.1: Collect user behavior data such as interaction frequency, task completion, and dwell time in the virtual game world, and perform denoising, normalization, and feature extraction on each set of collected behavior data. Then, create an analysis and adjustment model based on the input layer, bidirectional GRU layer, and output layer architecture.
[0022] S2.2: Divide the prepared data set into a training set and a test set, import the training set into the analysis and adjustment model, the model input layer passes the training set data into the bidirectional GRU layer, the bidirectional GRU layer performs forward GRU processing and reverse GRU processing on the training set data respectively, and sends the processing results to the output layer for post-processing output;
[0023] S2.3: The loss value between the model output result and the true value is calculated and analyzed through the cross entropy loss function, and the gradient of the loss value for each network layer of the model is calculated through the back propagation algorithm. The parameters of each network layer are optimized through the Adam optimizer. After each training cycle, the model is verified using the test set to evaluate its performance on unseen data. If the model performance does not meet the preset indicators, the model parameters are retrained and updated until the model performance meets the preset indicators.
[0024] S2.4: Input the user's latest behavioral characteristics into the trained analysis and adjustment model, predict the user's emotional state and game engagement in the virtual world through forward propagation, dynamically adjust the elements of the game environment and the difficulty of the game tasks based on the predicted emotional state and engagement, and then continuously monitor the behavior of new users in the adjusted environment and collect feedback to evaluate the effectiveness of the adjustment measures.
[0025] As a further solution of the present invention, the specific steps of real-time optimization of social network structure in step IV are as follows:
[0026] S3.1: Collect the social network topology structure inside the software and convert it into a corresponding graph structure, where the nodes of the graph represent users and the edges represent the connections between users. The corresponding adjacency matrix is determined based on the social attributes between users, multiple groups of individuals are randomly generated, and the initial positions of each group of individuals are set, where each group of individuals represents a potential optimization solution;
[0027] S3.2: Set an initial pheromone value for each edge in the graph structure. Then, each group of individuals starts from the initial position and moves in the social network. When selecting the next node, the selection probability of each node is calculated according to the pheromone concentration and the heuristic function, and the next node is selected by random selection based on the calculated selection probability.
[0028] S3.3: After each group of individuals completes a complete path iteration, the pheromone concentration of each edge in each path is updated according to the path quality constructed by each group of individuals, and node selection and pheromone update are repeated until the maximum number of iterations is reached or the pheromone concentration change reaches a preset threshold;
[0029] S3.4: Select the path with the highest pheromone concentration in all iterations as the optimal optimization solution, and optimize the structure of the social network accordingly according to the optimal optimization solution, evaluate the effect of the optimized social network, and continuously detect the interactive performance of the social network for subsequent updates and adjustments.
[0030] As a further solution of the present invention, the specific steps of designing diversified tasks for each user in step V are as follows:
[0031] S4.1: collect the types and parameters of tasks in the virtual game world, use the collected groups of tasks as nodes in the tree structure, and construct an initialization structure of the adjustment tree, where the root node represents the current task status and the child nodes represent executable tasks;
[0032] S4.2: Calculate the UCB value of each group of child nodes, and use the greedy strategy to select the child node with the highest UCB value layer by layer until the selected child node is a leaf node. Then stop selecting. If the state of the leaf node is not the terminal state, expand the node and add one or more child nodes.
[0033] S4.3: Perform random simulation on the expanded child nodes, execute the tasks and calculate the results, calculate the reward for completing the tasks by evaluating whether the tasks are successfully completed, update the number of visits and reward values of the nodes according to the simulation results, trace the simulation results back to the root node, and update the number of visits and reward values of the remaining nodes on the same path;
[0034] S4.4: Repeat the selection, expansion, simulation, and update steps until the maximum number of simulations is reached. After the tree structure is completed, the child node with the most visits and the highest reward value is selected as the optimal task based on the number of visits and rewards of the task corresponding to each child node. Then, the actual task setting is performed based on the selected optimal task, and user feedback is collected to evaluate the effectiveness of the task setting. The difficulty and participation of the task are adjusted based on the feedback.
[0035] As a further solution of the present invention, the specific steps of analyzing user feedback and interaction data in step VI are as follows:
[0036] S5.1: Collect users’ psychological counseling feedback and interaction data, pre-process the collected data, initialize the number of water cycles, the number of iterations, and the update rate of each cycle, and construct an initial solution set based on user feedback data to represent the quality of current counseling content and services;
[0037] S5.2: Initialize the evaporation rate and precipitation by random selection. In each iteration, randomly select a solution from the current solution set and randomly generate a random number from the range [0, 1]. If the random number is less than the evaporation rate, delete the solution from the current solution set. Then generate a new solution from the solution space by random selection or mutation of various heuristic generation methods.
[0038] S5.3: Combine the generated new solution with the current solution after evaporation to form a new solution set, then perform a quality assessment on each solution in the new solution set, select the optimal solution to enter the next round of water cycle based on the assessment result, and repeat the iterative update until the quality change of the solution is less than the preset threshold;
[0039] S5.4: The final solution obtained through iteration is converted into optimized consulting content and service quality. After the optimization plan is implemented, user feedback and interaction data are continuously collected, and the optimization effect is evaluated based on user satisfaction and consultation success rate.
[0040] Compared with the prior art, the present invention has the following beneficial effects:
[0041] 1. The intelligent virtual simulation method for promoting mental health of perinatal depressed women collects the internal social network topology of the software and converts it into a corresponding graph structure. It randomly generates multiple groups of individuals and sets the initial position of each group of individuals. Each group of individuals represents a potential optimization scheme. The initial pheromone value is set for each edge in the graph structure. After that, each group of individuals starts from the initial position and moves in the social network. When selecting the next node, the selection probability of each node is calculated according to the pheromone concentration and the heuristic function. The next node is selected by random selection based on the calculated selection probability. When each group of individuals completes a complete path iteration, the path quality constructed by each group of individuals is selected. Update the pheromone concentration of each edge in each path, repeat node selection and pheromone update until the maximum number of iterations is reached or the pheromone concentration change reaches the preset threshold, select the path with the highest pheromone concentration in all iterations as the optimal optimization plan, and optimize the social network structure accordingly according to the optimal optimization plan, evaluate the effect of the optimized social network, and continuously detect the interactive performance of the social network for subsequent updates and adjustments, which can effectively improve the quality and effectiveness of social connections, thereby providing each user with more and more helpful social support, improving the effectiveness of social interaction, and effectively improving the emotional recovery speed and social participation of perinatal women, and reducing loneliness.
[0042] 2. The intelligent virtual simulation method for promoting mental health of perinatal depressed women collects the types and parameters of tasks in the virtual game world, takes each group of collected tasks as nodes in the tree structure, constructs the initialization structure of the adjustment tree, the root node represents the current task state, the child nodes represent executable tasks, calculates the UCB value of each group of child nodes, and uses a greedy strategy to select the child node with the highest UCB value layer by layer until the selected child node is a leaf node, stops selecting, if the leaf node state is not a terminal state, expands the node, and adds one or more child nodes, performs random simulation on the expanded child nodes, executes tasks and calculates results, calculates the reward for completing the task by evaluating whether the task is successfully completed, updates the number of node visits and reward values according to the simulation results, and traces the simulation results back to the root Nodes are selected, and the number of visits and reward values of the remaining nodes on the same path are updated. The selection, expansion, simulation and update steps are repeated until the maximum number of simulations is reached. After the tree structure is completed, the child node with the most visits and the highest reward value is selected as the optimal task according to the number of visits and rewards of the tasks corresponding to each child node. Then, the actual task is set according to the selected optimal task, and user feedback is collected to evaluate the effectiveness of the task setting. The difficulty and participation of the task are adjusted according to the feedback, which can improve the user's enthusiasm and involvement, avoid the frustration or boredom caused by tasks that are too difficult or too easy, motivate users to continue to participate, enhance their sense of self-efficacy, and thus improve the effect of virtual simulation intervention, thereby enhancing the self-confidence and self-worth of women with peripartum depression, and helping with emotional regulation and psychological recovery. BRIEF DESCRIPTION OF THE DRAWINGS
[0043] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention.
[0044] Figure 1 This is a flowchart of an intelligent virtual simulation method for promoting mental health of women with perinatal depression proposed by the present invention. DETAILED DESCRIPTION
[0045] Example 1
[0046] Reference Figure 1 , an intelligent virtual simulation method for promoting mental health of women with perinatal depression, the specific steps of the promotion method are as follows:
[0047] To collect needs of women with perinatal depression and to establish a correlation between needs and software.
[0048] Specifically, we designed questionnaires covering mental health, social needs, emotional management, and professional support. We distributed the questionnaires online or face-to-face, then collected the completed questionnaires, conducted statistical analysis on the collected questionnaire data, identified the main mental health needs and concerns of users, conducted in-depth interviews with multiple groups of questionnaire participants through random sampling, recorded the participants' specific experiences and suggestions, organized the questionnaire and interview data into structured information, obtained the main needs and pain points of users, coded and classified the interview content, extracted entities from the organized demand data through natural language processing technology, including user demand keywords and functional module names, and determined the relationship between different entities based on the analysis results. Then, we used the Neo4j graph database to construct an association graph with the extracted entities as nodes and the corresponding relationships between nodes as edges. We analyzed the association graph through a graph convolutional network to explore the potential relationships between groups of nodes, and then visualized the constructed association graph through Graphviz. We also analyzed the structure of the graph, identified the relationship between core requirements and software modules, and regularly updated the association graph based on new demand information and changes in software modules.
[0049] Build a personalized virtual game world based on user needs.
[0050] Analyze and manage changes in user emotions and dynamically adjust the gaming environment.
[0051] Specifically, the user's behavioral data such as interaction frequency, task completion and residence time in the virtual game world are collected, and each group of collected behavioral data is denoised, normalized and feature extracted. Then, an analysis and adjustment model is created based on the input layer, bidirectional GRU layer and output layer architecture, and the prepared data set is divided into a training set and a test set. The training set is imported into the analysis and adjustment model. The input layer of the model passes the training set data into the bidirectional GRU layer, and the bidirectional GRU layer performs forward GRU processing and reverse GRU processing on the training set data respectively, and sends the processing results to the output layer for post-processing output. The loss value between the output result of the analysis and adjustment model and the true value is calculated through the cross entropy loss function, and the reverse GRU is used to calculate the loss value. The forward propagation algorithm calculates the gradient of the loss value for each network layer of the model, and optimizes the parameters of each network layer through the Adam optimizer. After each training cycle, the model is verified using the test set to evaluate its performance on unseen data. If the model performance does not meet the preset indicators, the model parameters are retrained and updated until the model performance reaches the preset indicators. The user's latest behavioral characteristics are input into the trained analysis and adjustment model, and the user's emotional state and game participation in the virtual world are predicted through forward propagation. According to the predicted emotional state and participation, the elements of the game environment and the difficulty of the game tasks are dynamically adjusted. Then, the behavior of new users in the adjusted environment is continuously monitored, and feedback is collected to evaluate the effectiveness of the adjustment measures.
[0052] Integrate social interaction functions for users to communicate and optimize social network structure in real time.
[0053] Specifically, the social network topology structure inside the software is collected and converted into a corresponding graph structure, where the nodes of the graph represent users and the edges represent the connections between users. The corresponding adjacency matrix is determined based on the social attributes between the users. Multiple groups of individuals are randomly generated and the initial positions of each group of individuals are set. Each group of individuals represents a potential optimization solution. An initial pheromone value is set for each edge in the graph structure. After that, each group of individuals starts from the initial position and moves in the social network. When selecting the next node, the selection probability of each node is calculated based on the pheromone concentration and the heuristic function. The next node is selected by random selection based on the calculated selection probability. When each group of individuals completes a complete path iteration, the pheromone concentration of each edge in each path is updated according to the path quality constructed by each group of individuals. Node selection and pheromone update are repeated until the maximum number of iterations is reached or the pheromone concentration change reaches a preset threshold. The path with the highest pheromone concentration is selected as the optimal optimization solution in all iterations. According to the optimal optimization solution, the social network is structurally optimized accordingly, the effect of the optimized social network is evaluated, and the interactive performance of the social network is continuously tested for subsequent updates and adjustments.
[0054] Example 2
[0055] Reference Figure 1 , an intelligent virtual simulation method for promoting mental health of women with perinatal depression, the specific steps of the promotion method are as follows:
[0056] Assess the engagement and difficulty of each task and design diverse tasks for each user.
[0057] Specifically, the types and parameters of tasks in the virtual game world are collected, and each group of collected tasks is used as a node in a tree structure to construct an initialization structure of the adjustment tree. The root node represents the current task state, and the child node represents the executable task. The UCB value of each group of child nodes is calculated, and the child node with the highest UCB value is selected layer by layer using a greedy strategy until the selected child node is a leaf node, and the selection is stopped. If the leaf node state is not a terminal state, the node is expanded and one or more child nodes are added. Random simulation is performed on the expanded child node, the task is executed and the result is calculated. By evaluating whether the task is successfully completed, the reward for completing the task is calculated, and the number of visits and reward values of the node are updated according to the simulation results. The simulation results are traced back to the root node, and the number of visits and reward values of the remaining nodes on the same path are updated. The selection, expansion, simulation and update steps are repeated until the maximum number of simulations is reached. After the tree structure is completed, the child node with the largest number of visits and the highest reward value is selected as the optimal task according to the number of visits and rewards of the task corresponding to each child node. Then, the actual task setting is performed according to the selected optimal task, and user feedback is collected to evaluate the effectiveness of the task setting, and the difficulty and participation of the task are adjusted according to the feedback.
[0058] Analyze user feedback and interaction data and provide professional psychological counseling.
[0059] Specifically, psychological counseling feedback and interaction data of users are collected, and the collected data are preprocessed to initialize the number of water cycles, the number of iterations, and the update rate of each cycle. An initial solution set is constructed according to user feedback data to represent the quality of current counseling content and services. The evaporation rate and precipitation are initialized by random selection. In each iteration, a solution is randomly selected from the current solution set, and a random number is randomly generated from the range of [0, 1]. If the random number is less than the evaporation rate, the solution is deleted from the current solution set. Then, new solutions are generated from the solution space by randomly selecting or mutating various heuristic generation methods. The generated new solutions are combined with the current solutions after evaporation to form a new solution set. After that, the quality of each solution in the new solution set is evaluated. According to the evaluation results, the optimal solution is selected to enter the next round of water cycle. The iterative update is repeated until the quality change of the solution is less than the preset threshold. The final solution obtained through iteration is the optimized counseling content and service quality. After the optimization plan is implemented, user feedback and interaction data are continuously collected, and the optimization effect is evaluated based on user satisfaction and consultation success rate.
[0060] Collect user feedback and usage data to evaluate software effectiveness and user satisfaction.
[0061] Continuously update software features and content based on user feedback and the latest research.
Claims
1. An intelligent virtual simulation method for promoting mental health of women with perinatal depression, characterized in that: The specific steps of the promotion method are as follows: Ⅰ. Collect the needs of women with perinatal depression and establish the connection between the needs and the software; Ⅱ. Build a personalized virtual game world based on user needs; Ⅲ. Analyze and manage user emotion changes and dynamically adjust the game environment; IV. Integrate social interaction functions to facilitate user communication and optimize the social network structure in real time; V. Evaluate the engagement and difficulty of each task and design diverse tasks for each user; VI. Analyze user feedback and interaction data and provide professional psychological counseling; Ⅶ. Collect user feedback and usage data to evaluate software effectiveness and user satisfaction; VIII. Continuously update software functions and content based on user feedback and latest research.
2. According to claim 1, an intelligent virtual simulation method for promoting mental health of women with perinatal depression is characterized in that: The specific steps for establishing the association between requirements and software described in step Ⅰ are as follows: S1.1: Design a questionnaire covering mental health, social needs, emotional management, and professional support, distribute the questionnaire online or in person, collect the completed questionnaires, conduct statistical analysis on the collected questionnaire data, and identify the main mental health needs and concerns of users; S1.2: Conduct in-depth interviews with multiple groups of questionnaire participants through random sampling, record the participants' specific experiences and suggestions, organize the questionnaire and interview data into structured information, obtain the main needs and pain points of users, and encode and classify the interview content; S1.3: Extract entities from the sorted demand data through natural language processing technology, including user demand keywords and function module names, and determine the relationship between different entities based on the analysis results. Then use the Neo4j graph database to construct a relationship graph with the extracted entities as nodes and the corresponding relationships between nodes as edges; S1.4: Analyze the association graph through graph convolutional networks to explore the potential relationships between groups of nodes, then visualize the constructed association graph through Graphviz, analyze the structure of the graph, identify the relationship between core requirements and software modules, and regularly update the association graph based on new requirement information and changes in software modules.
3. According to claim 2, an intelligent virtual simulation method for promoting mental health of women with perinatal depression is characterized in that: The specific steps for dynamically adjusting the game environment described in step III are as follows: S2.1: Collect user behavior data such as interaction frequency, task completion, and dwell time in the virtual game world, and perform denoising, normalization, and feature extraction on each set of collected behavior data. Then, create an analysis and adjustment model based on the input layer, bidirectional GRU layer, and output layer architecture. S2.2: Divide the prepared data set into a training set and a test set, import the training set into the analysis and adjustment model, the model input layer passes the training set data into the bidirectional GRU layer, the bidirectional GRU layer performs forward GRU processing and reverse GRU processing on the training set data respectively, and sends the processing results to the output layer for post-processing output; S2.3: The loss value between the model output result and the true value is calculated and analyzed through the cross entropy loss function, and the gradient of the loss value for each network layer of the model is calculated through the back propagation algorithm. The parameters of each network layer are optimized through the Adam optimizer. After each training cycle, the model is verified using the test set to evaluate its performance on unseen data. If the model performance does not meet the preset indicators, the model parameters are retrained and updated until the model performance meets the preset indicators. S2.4: Input the user's latest behavioral characteristics into the trained analysis and adjustment model, predict the user's emotional state and game engagement in the virtual world through forward propagation, dynamically adjust the elements of the game environment and the difficulty of the game tasks based on the predicted emotional state and engagement, and then continuously monitor the behavior of new users in the adjusted environment and collect feedback to evaluate the effectiveness of the adjustment measures.
4. According to claim 1, a method for promoting mental health of women with perinatal depression through intelligent virtual simulation, characterized in that: The specific steps of real-time optimization of social network structure described in step IV are as follows: S3.1: Collect the social network topology structure inside the software and convert it into a corresponding graph structure, where the nodes of the graph represent users and the edges represent the connections between users. The corresponding adjacency matrix is determined based on the social attributes between users, multiple groups of individuals are randomly generated, and the initial positions of each group of individuals are set, where each group of individuals represents a potential optimization solution; S3.2: Set an initial pheromone value for each edge in the graph structure. Then, each group of individuals starts from the initial position and moves in the social network. When selecting the next node, the selection probability of each node is calculated according to the pheromone concentration and the heuristic function, and the next node is selected by random selection based on the calculated selection probability. S3.3: After each group of individuals completes a complete path iteration, the pheromone concentration of each edge in each path is updated according to the path quality constructed by each group of individuals, and node selection and pheromone update are repeated until the maximum number of iterations is reached or the pheromone concentration change reaches a preset threshold; S3.4: Select the path with the highest pheromone concentration in all iterations as the optimal optimization solution, and optimize the structure of the social network accordingly according to the optimal optimization solution, evaluate the effect of the optimized social network, and continuously detect the interactive performance of the social network for subsequent updates and adjustments.
5. According to claim 3, an intelligent virtual simulation method for promoting mental health of women with perinatal depression is characterized in that: The specific steps for designing diversified tasks for each user described in step V are as follows: S4.1: collect the types and parameters of tasks in the virtual game world, use the collected groups of tasks as nodes in the tree structure, and construct an initialization structure of the adjustment tree, where the root node represents the current task status and the child nodes represent executable tasks; S4.2: Calculate the UCB value of each group of child nodes, and use the greedy strategy to select the child node with the highest UCB value layer by layer until the selected child node is a leaf node. Then stop selecting. If the state of the leaf node is not the terminal state, expand the node and add one or more child nodes. S4.3: Perform random simulation on the expanded child nodes, execute the tasks and calculate the results, calculate the reward for completing the tasks by evaluating whether the tasks are successfully completed, update the number of visits and reward values of the nodes according to the simulation results, trace the simulation results back to the root node, and update the number of visits and reward values of the remaining nodes on the same path; S4.4: Repeat the selection, expansion, simulation and update steps until the maximum number of simulations is reached. After the tree structure is completed, the child node with the most visits and the highest reward value is selected as the optimal task based on the number of visits and rewards of the task corresponding to each child node. Then, the actual task setting is performed based on the selected optimal task, and user feedback is collected to evaluate the effectiveness of the task setting. The difficulty and participation of the task are adjusted based on the feedback.
6. The intelligent virtual simulation method for promoting mental health of perinatal depressed women according to claim 5, characterized in that: The specific steps for analyzing user feedback and interaction data described in step VI are as follows: S5.1: Collect users’ psychological counseling feedback and interaction data, pre-process the collected data, initialize the number of water cycles, the number of iterations, and the update rate of each cycle, and construct an initial solution set based on user feedback data to represent the quality of current counseling content and services; S5.2: Initialize the evaporation rate and precipitation by random selection. In each iteration, randomly select a solution from the current solution set and randomly generate a random number from the range [0, 1]. If the random number is less than the evaporation rate, delete the solution from the current solution set. Then generate a new solution from the solution space by random selection or mutation of various heuristic generation methods. S5.3: Combine the generated new solution with the current solution after evaporation to form a new solution set, then perform a quality assessment on each solution in the new solution set, select the optimal solution to enter the next round of water cycle based on the assessment result, and repeat the iterative update until the quality change of the solution is less than the preset threshold; S5.4: The final solution obtained through iteration is converted into optimized consulting content and service quality. After the optimization plan is implemented, user feedback and interaction data are continuously collected, and the optimization effect is evaluated based on user satisfaction and consultation success rate.