Digital intelligence spirit recovery service platform and operation method thereof

Abstract

The invention provides a digital intelligence spirit recovery service platform and an operation method thereof, and the platform comprises an information collection module which collects user registration, behavior and scale data; the personalized evaluation module is used for generating a psychological evaluation report based on machine learning; the intelligent recommendation module formulates a rehabilitation plan including cognitive training, emotion support and social activities in combination with the report; the intelligent analysis module is used for analyzing the behavior data and providing a rehabilitation progress visual report; the virtual reality module supports emotion adjustment and cognitive training through an immersive scene; and the digital human module is used for providing emotion support and interactive guidance by means of natural language processing. All the modules work cooperatively, and individuation, precision and dynamic of mental rehabilitation service are achieved.

Description

Technical Field

[0001] The present application relates to the technical field of mental disorder rehabilitation, and in particular to a digital mental recovery service platform and its operation method. Background Art

[0002] Currently, existing rehabilitation models for mental disorders primarily focus on traditional medical interventions and psychotherapy. Service resources and personnel are unevenly distributed, resulting in low effective service coverage. Furthermore, many individuals experiencing recovery face difficulties integrating into society after treatment, lacking necessary social support and functional training. Therefore, it is crucial to develop a comprehensive platform for self-help and mutual assistance in mental recovery, leveraging modern technology for effective intervention. Summary of the Invention

[0003] This application provides a digital mental recovery service platform that integrates modern technology, optimizes service processes, and improves rehabilitation effects, providing high-quality support and intervention for people recovering from mental disorders in a systematic and long-term manner.

[0004] The specific plan of this application is as follows: A digital mental recovery service platform, including: Information collection module, used to collect user information, including user registration information, behavioral data and standardized scale data; The personalized assessment module uses machine learning algorithms to conduct personalized psychological assessments on users based on user information and generate assessment reports; An intelligent recommendation module, combining the assessment report with a recommendation algorithm, develops a personalized rehabilitation plan for the user, the rehabilitation plan including cognitive training, emotional support, and social activities; Intelligent analysis module, which uses machine learning technology to analyze user behavior data and provide visual reports on users' mood swings and recovery progress; Virtual reality module, which provides immersive scenes to support emotional regulation and cognitive training; The digital human module includes multiple preset characters and provides users with emotional support and interactive guidance through natural language processing technology.

[0005] Furthermore, the virtual reality module specifically includes: Design interactive scenarios and tasks to reconstruct cognitive models for users; Combine biofeedback technology with psychological intervention to monitor the user's emotional state in real time and dynamically adjust the cognitive training program.

[0006] By designing interactive scenarios and tasks to reconstruct cognitive models, immersive intervention can be provided for users' specific psychological states, helping them gradually establish positive cognitive logic in simulated real-life situations. Combining biofeedback technology with psychological intervention, real-time perception of users' emotional states and dynamic adaptation of training programs can be achieved, making cognitive training more targeted and timely, and improving the accuracy of psychological intervention and user participation.

[0007] Furthermore, the biofeedback technology is combined with psychological intervention to monitor the user's emotional state in real time and dynamically adjust the cognitive training program, specifically including: Real-time monitoring of user physiological data through heart rate sensors and skin galvanic response sensors; Automatically adjust the difficulty of the VR scene or provide guidance prompts based on the physiological data.

[0008] Using heart rate sensors and galvanic skin response sensors to collect physiological data in real time can objectively reflect the physiological manifestations of users' emotional fluctuations, avoiding the bias of relying solely on subjective reports; automatically adjusting the difficulty of VR scenes or providing guidance prompts based on physiological data can realize a closed-loop intervention model of "monitoring-feedback-adjustment", ensuring that the training intensity matches the user's current psychological tolerance, reducing the user's psychological burden while ensuring the training effect.

[0009] Furthermore, the personalized assessment module specifically includes: Data preprocessing unit, used to preprocess user data, including filling missing values with the mean or mode, and detecting and correcting outliers using the Z-score method; A feature extraction unit is used to extract key features from user data, including emotional state features, sociability features, and life satisfaction features; Feature selection unit, which uses Pearson correlation coefficient to select relevant features from key features; A model building unit, which builds and trains a linear regression model based on the relevant features; The result calculation unit generates evaluation results based on the linear regression model, including emotional state scores, social ability scores and life satisfaction scores.

[0010] The data preprocessing unit improves the quality of original data and avoids noise interference through mean filling, mode filling and outlier correction; the feature extraction and selection unit focuses on core dimensions such as emotions, sociality, and life satisfaction, and uses the Pearson correlation coefficient to screen key features and reduce the influence of irrelevant variables; quantitative scores are generated based on linear regression models to make psychological assessment results more scientific and operational, providing accurate data support for the formulation of personalized rehabilitation plans.

[0011] Furthermore, the platform also includes a dynamic adjustment module for regularly updating the rehabilitation plan based on user behavior data and evaluation reports, specifically including: Regularly obtain user behavior data and evaluation reports; Adjust training intensity and content based on user recovery progress; Optimize training methods based on user engagement and effectiveness; Dynamically adjust the frequency of emotional support based on mood swings.

[0012] The dynamic adjustment module regularly obtains user behavior data and evaluation reports to track rehabilitation progress in real time and avoid "one-size-fits-all" intervention; it adjusts training intensity and content based on rehabilitation progress and optimizes training methods based on participation, ensuring that the rehabilitation plan dynamically iterates with the user's status and maintains a balance between user enthusiasm and challenge; it adjusts the frequency of emotional support based on emotional fluctuations to achieve "precise response" to users' psychological needs and enhance the timeliness and effectiveness of emotional support.

[0013] Furthermore, the use of machine learning technology to analyze user behavior data specifically includes: Clean user behavior data and standardize scale scores; Extract sentiment keywords and activity frequencies from the cleaned data; Use time series analysis models to predict user sentiment fluctuation trends; Clustering algorithms are used to divide users into groups at different rehabilitation stages.

[0014] Data cleaning and standardization improve the quality of behavioral data and lay the foundation for subsequent analysis; emotional keywords and activity frequency extraction can explore users' potential psychological states and behavioral patterns, and time series analysis models predict emotional fluctuation trends to help identify psychological risk points in advance; clustering algorithms divide rehabilitation stage groups, supporting the platform to provide differentiated intervention strategies for users at different stages, achieving optimized resource allocation and precise services.

[0015] Furthermore, the recommendation algorithm adopts at least one of the following algorithms: Collaborative filtering algorithm recommends appropriate cognitive training tasks and emotional support content based on the rehabilitation plans of similar users; Content recommendation algorithm, based on user evaluation tags, matching predefined rehabilitation resource libraries; Reinforcement learning algorithm dynamically optimizes recommendation strategies to adapt to user feedback.

[0016] The collaborative filtering algorithm leverages the rehabilitation experience of similar users to recommend proven and effective solutions to new users, reducing trial and error costs. The content recommendation algorithm matches predefined resource libraries based on evaluation tags to ensure that the recommended content is highly consistent with the user's psychological needs. The reinforcement learning algorithm dynamically optimizes the recommendation strategy and continuously iterates the recommendation model through user feedback, allowing the rehabilitation plan to evolve dynamically with individual user differences and real-time status, thereby improving the intelligence level of the recommendation system.

[0017] Furthermore, the platform also includes a privacy protection module, which is specifically used to: Encrypt user data transmission via SSL / TLS protocol; Set up multi-level access control to limit access to sensitive data; Allow users to choose the scope of information sharing.

[0018] The SSL / TLS protocol encrypts data transmission, preventing user information from being stolen or tampered with during network transmission and ensuring data integrity; multi-level access control and sensitive data access restrictions build a data security barrier from the architectural level; users can independently choose the scope of information sharing, giving users control over their personal data, enhancing user trust in the platform, and at the same time complying with privacy protection regulations and reducing data security risks.

[0019] Furthermore, the present application also provides a method for operating a digital mental recovery service platform, comprising the following steps: S1, users register and complete an initial psychological assessment; S2, generates a personalized rehabilitation plan and initiates real-time data monitoring; S3, dynamically adjusts the rehabilitation plan based on user behavior data and physiological feedback; S4, providing multi-dimensional psychological support through digital human interaction and VR training; S5, regularly generate visual rehabilitation progress reports.

[0020] The operation method achieves full-link digital coverage from user access to rehabilitation management through a closed-loop process of "registration-assessment-plan generation-dynamic adjustment-multi-dimensional support-effect feedback"; real-time data monitoring and physiological feedback-driven plan adjustment ensure that the intervention plan always meets the dynamic needs of users; the combination of digital human interaction and VR training provides multimodal psychological support to meet the interaction preferences of different users; visual rehabilitation reports help users and professionals intuitively grasp the progress, improving the transparency of the rehabilitation process and user participation.

[0021] Compared with the existing technology, the beneficial effects of this application are as follows: this digital mental recovery service platform constructs a full-chain digital psychological rehabilitation system through the collaboration of multiple modules: the information collection module integrates multi-source data to lay a personalized foundation; the personalized assessment module uses machine learning to achieve quantitative psychological portraits; the intelligent recommendation module dynamically generates a three-dimensional rehabilitation plan based on the assessment results, including cognitive training, emotional support, and social activities; the intelligent analysis module tracks the rehabilitation progress in real time through behavioral data analysis and visualization reports; the virtual reality module provides immersive scene intervention; and the digital human module enhances emotional companionship through natural language interaction. The modules are interconnected in data and complementary in function, forming a closed-loop service system from assessment to intervention, from monitoring to feedback, breaking through the limitations of traditional psychological rehabilitation services that are highly subjective and inefficient, and realizing digital upgrades with precise diagnosis, personalized treatment, and dynamic adjustment, significantly improving the scientific nature, effectiveness, and user experience of mental rehabilitation services. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the embodiments of the present drawings or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present drawings. For ordinary technicians in this field, other drawings can be obtained based on the structures shown in these drawings without paying any creative work.

[0023] Figure 1 This is a schematic diagram of the platform principle of this application. DETAILED DESCRIPTION

[0024] In order to enable those skilled in the art to better understand the technical solution, the technical solution is described in detail below in conjunction with the embodiments. The description in this section is only exemplary and explanatory and should not have any limiting effect on the scope of protection of this application.

[0025] Example 1: This application provides a digital mental recovery service platform that integrates Chinese culture and social environment, adopts the Chinese recovery model, emphasizes individual functional recovery in the community, and develops rehabilitation plans by enhancing self-management ability and social participation. It includes the following core elements: (1) Rebuilding the user’s connection with reality through cognitive training; (2) Enhance users’ social integration through regular daily activities and social interactions; (3) Cultivate resilience and positive self-positioning based on user strengths analysis; (4) Dynamically adjust the rehabilitation plan to match the recovery needs of users at different stages.

[0026] The platform adopts the following two architectural designs: Cloud computing architecture: The platform uses cloud computing technology to build a unified back-end service, separate data storage, computing and application logic, and ensure the scalability and high availability of the platform.

[0027] Microservice architecture: Using microservice architecture design, core functional modules (such as user management, data analysis, social interaction, etc.) are split into independent services to improve the flexibility of system maintenance and development.

[0028] like Figure 1 As shown in the figure, a digital mental recovery service platform includes the following main modules: The information collection module is used to collect user information, including user registration information, behavioral data and standardized scale data.

[0029] Specifically, the information collection module mainly collects the following user information: User registration information: including age, gender, medical history, medication usage, etc.

[0030] Standardized scale data: including the Assessment of Affective States Scale (PANAS), Satisfaction with Life Scale (SWLS), Social Support Rating Scale (SSRS), etc.

[0031] Behavioral data: user interaction records on the platform, including emotional diaries, cognitive training completion status, social interaction frequency, etc.

[0032] After data collection, it is represented in matrix form (X), where each row of the matrix corresponds to a user and each column corresponds to a feature (such as PANAS score, SWLS score, interaction frequency, etc.).

[0033] ; Here, m represents the number of users and n represents the number of features.

[0034] The personalized assessment module uses machine learning algorithms based on user information to conduct personalized psychological assessments of users and generate assessment reports.

[0035] Specifically, the personalized assessment module includes the following units: The data preprocessing unit is used to preprocess user data, including filling missing values with the mean or mode, and detecting and correcting outliers through the Z-score method.

[0036] (1) Missing value processing: Mean filling: For numerical data (such as scale scores), fill missing values with the mean of the feature: ; Mode filling: For categorical data (such as gender), fill with the mode of the feature.

[0037] (2) Outlier detection: Z-score method: Calculate the standard score Z for each data point. If the Z-score exceeds the first threshold, it is considered an outlier: ; in, and They represent the mean and standard deviation of the j-th column feature respectively.

[0038] Delete or modify: Delete or modify outliers based on business needs.

[0039] (3) Data standardization: Min-Max Normalization: Scale the data to the interval [0,1]: ; in, Represents the data of feature j of user i after normalization; represents the minimum eigenvalue; represents the maximum eigenvalue.

[0040] The feature extraction unit is used to extract key features from user data, including emotional state features, sociability features, and life satisfaction features.

[0041] Emotional state characteristics: Positive emotion scores (PA) and negative emotion scores (NA) were extracted from the PANAS scale; Social competence characteristics: Social support scores were extracted from the SSRS scale.

[0042] Life satisfaction characteristics: The overall life satisfaction score was extracted from the SWLS scale.

[0043] The feature selection unit uses the Pearson correlation coefficient to filter out relevant features from key features.

[0044] Specifically, the Pearson correlation coefficient is used to analyze the correlation between the feature and the target variable, and the relevant features with correlation values greater than the second threshold are screened out. The calculation formula is as follows: ; in, represents the eigenvalue; Represents the target vector, such as rehabilitation effect score; represents the mean of the eigenvalues; represents the mean of the target vector.

[0045] A model building unit builds and trains a linear regression model based on the relevant features.

[0046] (1) Constructing a linear regression model: ; Among them, y represents the rehabilitation effect score; represents the nth feature; represents the regression coefficient; represents the error term.

[0047] (2) Model training and validation: Divide the dataset into training and testing sets; The regression coefficients were estimated using the least squares method; The mean squared error (MSE) was used to evaluate the model performance.

[0048] The result calculation unit generates evaluation results based on the linear regression model, including emotional state scores, social ability scores and life satisfaction scores.

[0049] The intelligent recommendation module combines the assessment report and uses a recommendation algorithm to develop a personalized rehabilitation plan for the user, where the rehabilitation plan includes cognitive training, emotional support, and social activities.

[0050] Specifically, the recommendation algorithm adopts at least one of the following algorithms: Collaborative filtering algorithm recommends appropriate cognitive training tasks and emotional support content based on the rehabilitation plans of similar users; Content recommendation algorithm, based on user evaluation tags, matching predefined rehabilitation resource libraries; Reinforcement learning algorithms dynamically optimize recommendation strategies to adapt to user feedback. For example, if a user's score improves after completing a task, the weight of similar tasks will be increased.

[0051] Collaborative filtering algorithms are mainly divided into user-based collaborative filtering and item-based collaborative filtering. The specific steps and processes are as follows: 1) Data preparation: Collect user historical behavior data, including completed rehabilitation plans, task scores, etc.

[0052] This data constitutes a user-task matrix, where each user has different ratings for different tasks.

[0053] 2) Similarity calculation: Use cosine similarity to calculate the similarity between different users.

[0054] For example, for user A and user B, their similarity is determined by calculating their ratings for the same task.

[0055] 3) Neighbor user selection: Based on the similarity, users with a value higher than a preset threshold are selected as neighbors of the target user.

[0056] 4) Recommendation Generation: Among the rehabilitation tasks completed by neighbor users, filter out the tasks that the target user has not participated in.

[0057] According to the ratings of neighbor users on these tasks, a weighted average is taken to generate a recommendation list.

[0058] 5) Optimization and feedback: Collect feedback from target users on recommendation tasks and optimize user similarity models and recommendation strategies based on feedback data.

[0059] The content recommendation algorithm is based on matching user evaluation tags with the content of the rehabilitation resource library. The specific steps and process are as follows: 1) User evaluation label collection: Collect evaluation labels such as users' mental state, symptom description, interests and hobbies, etc.

[0060] 2) Construction of rehabilitation resource library: Create a resource library that includes various rehabilitation tasks, cognitive training courses, etc., and mark them with applicable tags.

[0061] 3) Tag matching: Compare user tags with task tags in the resource library and filter out tasks with high tag matching.

[0062] 4) Task sorting and recommendation: Tasks are sorted based on factors such as label matching, task effect historical data, and other factors, and the top-ranked tasks are recommended to users.

[0063] 5) User feedback integration: Collect user participation and effectiveness feedback on recommendation tasks, and adjust resource library labels and recommendation algorithms.

[0064] Reinforcement learning mainly explores and optimizes recommendation strategies through trial and error. The specific steps and process are as follows: 1) Environmental modeling: The user status and feedback are regarded as the environment, the recommendation strategy is regarded as the agent, and the recommendation result is regarded as the action.

[0065] 2) Status representation: The user’s current usage data, behavior history, actual rehabilitation tasks participated in, and emotional state are used as state inputs.

[0066] 3) Strategy training: Use deep reinforcement learning methods to build and train intelligent agents so that they can learn to choose the best task in different states.

[0067] The reward function is defined based on the degree of improvement in the user's score after completion.

[0068] 4) Strategy Update: Collect changes in user task ratings and update rewards and strategies.

[0069] 5) Dynamic Optimization: Continuously adjust strategy parameters through user feedback and environmental changes to ensure the best match between the rehabilitation plan and the user's current status.

[0070] These three algorithms use different paths to make personalized task recommendations, and their combined use can maximize the adaptability and effectiveness of rehabilitation programs.

[0071] The intelligent analysis module uses machine learning technology to analyze user behavior data and provide visual reports on users' emotional fluctuations and recovery progress.

[0072] Specifically, using machine learning technology to analyze user behavior data includes: Clean user behavior data and standardize scale scores, such as invalid entries in mood diaries; Extract emotional keywords (e.g., “stress,” “loneliness”) and activity frequency (e.g., number of social interactions per day) from the cleaned data; Use a time series analysis model to predict user mood fluctuation trends, such as the risk of depression in the next week. The time series analysis model can use the LSTM long short-term memory network model. The specific process is as follows: 1) Data preparation: Collect time series data, including users' emotional diaries, social interaction frequencies, physiological data (such as heart rate changes), etc. The data should be arranged in chronological order and missing values should be filled.

[0073] 2) Feature Engineering: Extract features such as the frequency of emotional keywords, emotional state scores, and sociability characteristics, and standardize the features to make them suitable for model training.

[0074] 3) Model construction and training: Build a model using LSTM, which is suitable for processing time series data and capturing long-term and short-term dependencies. Divide the data into training and test sets, train the LSTM model using the training set, and optimize the parameters to minimize the prediction error.

[0075] 4) Prediction and evaluation: The model is fed with recent data to predict mood swings or depression risk within the next week. The model's prediction accuracy is evaluated based on the test set, and parameters are adjusted to improve model performance.

[0076] 5) Deployment and application: Deploy the model to a real-time system, regularly update the prediction results, and use the prediction results to assist users or administrators in timely psychological intervention.

[0077] Use clustering algorithms to divide users into groups at different recovery stages, such as "initial adaptation period" and "stable recovery period". The clustering algorithm can use the K-means algorithm. The specific process is as follows: 1) Data preparation: Collect multi-dimensional data during the user's rehabilitation process, including emotional state scores, number of activities participated in, physiological feedback, etc., and clean and standardize the data according to business needs.

[0078] 2) Feature Selection: Select important characteristics that affect the recovery stage, such as social skills, emotional stability, and activity participation.

[0079] 3) K-means algorithm execution: Determine the number of clusters k and the number of rehabilitation stages to be divided (for example, three stages: initial adaptation period, stable recovery period, and recovery growth period). Use the K-means algorithm to cluster user feature data and assign users to different rehabilitation stage groups.

[0080] 4) Result analysis and application: Analyze the characteristics of each group, determine the group's position and needs in the rehabilitation cycle, design differentiated intervention strategies and resource allocation plans for different groups, and achieve personalized rehabilitation.

[0081] 5) Dynamic adjustment and optimization: Regularly re-cluster based on new data, adjust group divisions, and use clustering results to optimize resource allocation and rehabilitation program adaptability.

[0082] By applying LSTM and K-means to user emotional trend prediction and rehabilitation stage division, the platform can provide more accurate emotional monitoring and personalized rehabilitation plans, striving to quickly respond to changes in users' mental state and provide timely and effective psychological support.

[0083] Visual reports include: Emotional fluctuation trend chart, showing how the user's emotional state changes over time; The recovery progress radar chart displays multi-dimensional assessment results, including emotional state, social skills, life satisfaction, etc.

[0084] Virtual reality module provides immersive scenes to support emotional regulation and cognitive training.

[0085] Specifically, the virtual reality module reconstructs cognitive patterns for users by designing interactive scenarios and tasks, improving their emotional management and coping abilities; it combines biofeedback technology with psychological intervention to monitor users' emotional states in real time and dynamically adjust cognitive training programs.

[0086] Specifically, the user's physiological data is monitored in real time through a heart rate sensor and a skin galvanic response sensor; the difficulty of the VR scene is automatically adjusted or guidance prompts are provided based on the physiological data.

[0087] Users wear VR devices to participate in various interactive scenarios and tasks. Biofeedback technology uses sensors (such as heart rate sensors and galvanic skin response sensors) to collect real-time physiological data from users. This physiological data reflects the user's emotional state. For example, an increased heart rate or increased galvanic skin response may indicate that the user is nervous or anxious.

[0088] Psychological intervention content is pre-designed within VR scenarios. For example, when users are engaging in social skills training, virtual characters interact with them, guiding them in learning and practicing social skills. If biofeedback technology detects a user's emotional state, the system automatically adjusts the difficulty of tasks within the VR scenario or provides appropriate guidance, such as reducing the complexity of social interactions or providing more social skills guidance, to dynamically adjust the cognitive training program.

[0089] The digital human module includes multiple preset characters and provides users with emotional support and interactive guidance through natural language processing technology.

[0090] A time elf character, used to remind users to carry out daily rehabilitation activities and time management; The Soul Explorer role provides emotional support and psychological guidance through natural language processing technology; The role of a life wise man provides life advice and rehabilitation guidance based on user data analysis.

[0091] Using speech recognition and natural language processing technologies, the Digital Human understands the user's voice or text input and analyzes their emotional state and needs. Based on this analysis, the Digital Human selects an appropriate response from pre-set dialogue templates and a knowledge base, providing feedback to the user in the form of voice or text. For example, if a user expresses anxiety, the Mind Explorer will respond with soothing words and guide the user through relaxation exercises.

[0092] The platform also includes a dynamic adjustment module, which is used to regularly update the rehabilitation plan based on user behavior data and evaluation reports to ensure that the rehabilitation plan is real-time and effective.

[0093] The dynamic adjustment module is a key part to ensure the real-time effectiveness and adaptability of the rehabilitation plan. This process can be achieved through a variety of algorithms, one of which is the reinforcement learning (RL) algorithm. The specific steps are as follows: 1) Data collection and preprocessing: Continuously collect users' real-time behavior data, emotional state changes, completed tasks, feedback ratings, etc., and clean and standardize the collected data to ensure the quality of the input data.

[0094] 2) State modeling: The user's current state (including emotions, behavioral data, and historically completed rehabilitation plans) is modeled as an environmental state, and emotional state scores, task completion, engagement, etc. are used as state feature inputs.

[0095] 3) Application of reinforcement learning algorithm: Using DQN (Deep Q Network) or other deep reinforcement learning algorithms, each possible adjustment (such as adding social activities, replacing training tasks) is considered an action.

[0096] 4) Reward function design: The change in the user's feedback score after completing the task is regarded as an immediate reward, and the long-term improvement in the user's mental health score is regarded as a long-term reward.

[0097] 5) Strategy training and optimization: Through trial and error, the RL agent learns to choose the action that maximizes the reward in a given state, and uses the Q-learning algorithm to update the Q value to gradually find the optimized strategy.

[0098] 6) Dynamically adjust execution: Based on the latest feedback data from users, the intelligent agent adjusts rehabilitation training in real time, striking a balance between testing new strategies and utilizing existing best strategies to ensure optimization.

[0099] 7) Model evaluation and update: Regularly evaluate the performance of the algorithm used for dynamic adjustment, and update the strategy model in a timely manner based on the evaluation results through indicators such as the degree of improvement in user emotions and the rate of achievement of standards, so as to continuously improve the adaptability of the rehabilitation plan.

[0100] Specifically, we regularly collect user behavior data and assessment reports, and adjust training intensity and content based on the user's recovery progress. For example, if the assessment report shows that the user has made significant progress in a certain area (such as social skills), the rehabilitation plan can appropriately reduce the intensity of training for that area and increase the content and time of training for other relatively weak areas (such as cognitive ability).

[0101] Optimize training methods based on user engagement and effectiveness, and dynamically adjust the frequency of emotional support based on mood swings. For example, if behavioral data reveals low user engagement and poor results in a certain type of training activity (such as a specific cognitive training task), the difficulty of the training task can be reduced, or a more suitable training method can be used. At the same time, adjust the method and frequency of emotional support based on user mood swings. For example, when a user's mood fluctuates significantly, increase the frequency of digital human companionship and communication, providing more emotionally comforting content.

[0102] The platform also includes a community interaction module that supports the following functions: Users share their life and work experiences and participate in themed discussion groups; Push relevant activity information based on content recommendation algorithms to promote mutual assistance and co-creation among users; Provide privacy setting options to allow users to choose the scope of information sharing.

[0103] The entire platform uses SSL / TLS protocol for data transmission encryption to ensure the security of user information. At the same time, user data is stored in a protected database with strict access rights.

[0104] The platform provides users with information sharing selection functions, allowing them to independently choose the scope of information sharing and protect user data privacy.

[0105] When it comes to providing users with information sharing options and protecting their data privacy, the following approaches are used to provide users with greater security and control.

[0106] 1) Blockchain technology integration: Recording user data access and change history on the blockchain makes data access transparent and tamper-proof, allowing users to view and track all access records.

[0107] 2) Data anonymization: User data is anonymized and differential privacy technology is used to ensure that even if the data is analyzed or mined, the identity of individual users cannot be identified.

[0108] 3) Real-time privacy dashboard: Provides a visual privacy dashboard that allows users to see in real time what data is being shared and with whom, and can modify their privacy settings instantly.

[0109] 4) Regular privacy reviews and reports: Conduct regular privacy audits and generate reports to show users how the platform protects their data privacy and the data access status due to changes in user settings.

[0110] 5) Dynamic adjustment of privacy protocols: Allow users to dynamically adjust their privacy protocols and sharing conditions when applying different functions on the platform, and set customized privacy levels based on the function type.

[0111] 6) Intelligent analysis of access logs: Use intelligent analysis tools to help users understand data access logs, identify any suspicious access patterns or potential privacy risks, and issue automatic alerts.

[0112] Example 2: This application is based on the digital mental recovery service platform in Example 1 and further provides a method for operating the digital mental recovery service platform, including the following steps: S1, users register and complete an initial psychological assessment.

[0113] S2, generates a personalized rehabilitation plan and starts real-time data monitoring.

[0114] S3, dynamically adjusts the rehabilitation plan based on user behavior data and physiological feedback.

[0115] S4, provides multi-dimensional psychological support through digital human interaction and VR training.

[0116] S5, regularly generate visual rehabilitation progress reports.

[0117] The specific applications are as follows: In a hospital setting, a patient with depression completed a positive and negative emotion scale assessment using the platform. The platform determined they had "high negative emotions and low positive emotions" and developed a plan for them, including daily mindfulness meditation and VR emotional relaxation training. The digital human guided the user in keeping a daily emotional diary and adjusted the difficulty of the VR scenes through biofeedback. After one month, the user's emotional stability score increased by 30%, and the frequency of social interactions increased by 50%.

[0118] Community scenarios: Community rehabilitation centers deploy a VR social training system to help individuals with social anxiety simulate job interviews. The platform dynamically adjusts the interviewer's facial expressions and question intensity based on the user's heart rate data, gradually reducing anxiety levels. The platform also recommends users join a "workplace skills support group" to boost confidence through community sharing.

[0119] Example 3: Establish an AI-based self-assessment and feedback system to enable users to assess their own conditions in daily life and receive timely feedback and guidance.

[0120] The specific implementation plan is as follows: Chatbot: Introduce AI chatbots to conduct self-assessment of emotional status, allowing users to understand their own emotions and needs through conversations.

[0121] Real-time feedback mechanism: Generates targeted suggestions and coping strategies based on user input of emotions and experiences, such as deep breathing exercises and relaxation techniques; Self-monitoring tool: Integrated mood diary function, users can record and review personal mood changes, thereby improving emotion recognition ability and self-awareness.

[0122] Data analysis platform: Generates visual reports based on user emotion records to help users identify triggers and improvement directions.

[0123] It should be noted that the present application is not limited to the above-mentioned embodiments. The above-mentioned embodiments are merely examples, and any embodiments having substantially the same structure and effect as the technical concept within the scope of the present application are all included in the technical scope of the present application. In addition, without departing from the scope of the present application, any other embodiments that can be conceived by those skilled in the art and that combine some of the constituent elements in the embodiments are also included in the scope of the present application.

Claims

1. A digital mental recovery service platform, characterized by: include: Information collection module, used to collect user information, including user registration information, behavioral data and standardized scale data; The personalized assessment module uses machine learning algorithms to conduct personalized psychological assessments on users based on user information and generate assessment reports; An intelligent recommendation module, combining the assessment report with a recommendation algorithm, develops a personalized rehabilitation plan for the user, the rehabilitation plan including cognitive training, emotional support, and social activities; The intelligent analysis module uses machine learning technology to analyze user behavior data and provide visual reports on users' mood swings and recovery progress. The use of machine learning technology to analyze user behavior data specifically includes: Clean user behavior data and standardize scale scores; Extract sentiment keywords and activity frequencies from the cleaned data; Use time series analysis models to predict user sentiment fluctuation trends; Use clustering algorithms to divide users into groups at different recovery stages; Virtual reality module, which provides immersive scenes to support emotional regulation and cognitive training; The digital human module includes multiple preset characters and provides users with emotional support and interactive guidance through natural language processing technology.

2. The digital mental recovery service platform according to claim 1, characterized in that: The virtual reality module specifically includes: Design interactive scenarios and tasks to reconstruct cognitive models for users; Combine biofeedback technology with psychological intervention to monitor the user's emotional state in real time and dynamically adjust the cognitive training program.

3. The digital mental recovery service platform according to claim 2, characterized in that: The method combines biofeedback technology with psychological intervention to monitor the user's emotional state in real time and dynamically adjust the cognitive training program, specifically including: Real-time monitoring of user physiological data through heart rate sensors and skin galvanic response sensors; Automatically adjust the difficulty of the VR scene or provide guidance prompts based on the physiological data.

4. The digital mental recovery service platform according to claim 1, characterized in that: The personalized assessment module specifically includes: Data preprocessing unit, used to preprocess user data, including filling missing values with the mean or mode, and detecting and correcting outliers using the Z-score method; A feature extraction unit is used to extract key features from user data, including emotional state features, sociability features, and life satisfaction features; Feature selection unit, which uses Pearson correlation coefficient to select relevant features from key features; A model building unit, which builds and trains a linear regression model based on the relevant features; The result calculation unit generates evaluation results based on the linear regression model, including emotional state scores, social ability scores and life satisfaction scores.

5. The digital mental recovery service platform according to claim 1, characterized in that: The platform also includes a dynamic adjustment module for regularly updating the rehabilitation plan based on user behavior data and assessment reports, specifically including: Regularly obtain user behavior data and evaluation reports; Adjust training intensity and content based on user recovery progress; Optimize training methods based on user engagement and effectiveness; Dynamically adjust the frequency of emotional support based on mood swings.

6. The digital mental recovery service platform according to claim 1, characterized in that: The recommendation algorithm adopts at least one of the following algorithms: Collaborative filtering algorithm recommends appropriate cognitive training tasks and emotional support content based on the rehabilitation plans of similar users; Content recommendation algorithm, based on user evaluation tags, matching predefined rehabilitation resource libraries; Reinforcement learning algorithm dynamically optimizes recommendation strategies to adapt to user feedback.

7. The digital mental recovery service platform according to claim 1, characterized in that: The platform also includes a privacy protection module, which is specifically used to: Encrypt user data transmission via SSL / TLS protocol; Set up multi-level access control to limit access to sensitive data; Allow users to choose the scope of information sharing.

8. A method for operating the digital mental recovery service platform according to any one of claims 1 to 7, characterized in that: The following steps are involved: S1, users register and complete an initial psychological assessment; S2, generates a personalized rehabilitation plan and initiates real-time data monitoring; S3, dynamically adjusts the rehabilitation plan based on user behavior data and physiological feedback; S4, providing multi-dimensional psychological support through digital human interaction and VR training; S5, regularly generate visual rehabilitation progress reports.

Images