Personalized mental health assessment and intervention method based on AI

Abstract

The invention relates to the technical field of medical care, in particular to an AI-based personalized mental health assessment and intervention method, which comprises the following steps: S1, data acquisition: acquiring multi-dimensional data of a user; s2, data preprocessing: preprocessing the collected multi-dimensional data; s3, psychological health portrait construction: generating a personalized psychological health portrait of the user; s4, social relation psychological influence analysis: identifying an emotion change mode, a negative emotion propagation path and a social isolation risk of the user in social interaction, and dynamically analyzing emotion propagation characteristics in a social network structure and potential influence factors of a user psychological state; s5, making a personalized psychological health intervention scheme: generating the personalized psychological health intervention scheme; the method disclosed by the invention not only reveals the potential influence of social interaction on mental health, but also provides a data-driven scientific basis for further personalized mental health intervention.

Description

Technical Field

[0001] The present invention relates to the field of medical care technology, and in particular to an AI-based personalized mental health assessment and intervention method. Background Art

[0002] With the increase of social pressure and the accelerated pace of life, mental health issues are becoming a hot topic of global concern. Mental health assessment and intervention are of great significance in preventing mental illness and improving the quality of life. In recent years, the development of artificial intelligence technology has provided new ideas for mental health management. By integrating multidimensional data, the user's mental state can be captured more comprehensively. At the same time, the profound impact of social relationships on mental health has gradually attracted the attention of researchers. For example, the spread of negative emotions may aggravate individual psychological problems, and the risk of isolation may lead to negative states such as depression and anxiety. Therefore, how to achieve accurate mental health assessment and intervention based on multidimensional data and social relationships has become an important issue that needs to be solved urgently.

[0003] Existing mental health assessment methods mostly focus on single-dimensional data analysis, such as psychological state assessment based on questionnaires or emotion recognition based on physiological signals. This method cannot fully reflect the user's mental health status. In addition, most assessment methods lack in-depth analysis of users' social relationships and cannot reveal the potential impact of emotional transmission and social isolation on mental health. More importantly, traditional methods find it difficult to dynamically adjust intervention plans and cannot adapt to changes in users' real-time psychological states, resulting in insufficient accuracy and effectiveness of intervention measures. Existing technologies still have obvious limitations in data processing efficiency, assessment accuracy and intervention personalization.

[0004] The present invention aims to provide an AI-based personalized mental health assessment and intervention method to improve the real-time and scientific nature of intervention, thereby achieving accurate, intelligent, and personalized mental health management and meeting the increasingly diverse mental health needs of users. Summary of the invention

[0005] The present invention provides an AI-based personalized mental health assessment and intervention method.

[0006] The AI-based personalized mental health assessment and intervention method includes the following steps: S1, data collection: collect multi-dimensional data of users, including physiological data (heart rate, respiratory rate, skin galvanic response), behavioral data (activity trajectory, exercise volume), psychological state data (emotional log, sleep quality) and social interaction data (social media activities, communication frequency, social network structure); S2, data preprocessing: preprocess the collected multi-dimensional data, including cleaning, denoising, missing value filling and standardization; S3, Mental Health Portrait Construction: Based on the preprocessed multi-dimensional data, generate a personalized mental health portrait of the user, including the user's mental health status score, mental health score, and intervention needs; S4, Analysis of the Psychological Impact of Social Relationships: By constructing the user's social relationship network, identify the emotional change patterns, negative emotion propagation paths, and social isolation risks in the user's social interactions, dynamically analyze the emotional propagation characteristics in the social network structure and the potential influencing factors of the user's mental state, and generate an impact assessment report on the influence of social relationships on mental health; S5, Development of a Personalized Mental Health Intervention Plan: According to the user's mental health portrait and the results of the analysis of the psychological impact of social relationships, generate a personalized mental health intervention plan, including behavior guidance, emotion management suggestions, social activity recommendations, and psychological counseling suggestions.

[0007] Optionally, the data collection in S1 includes: S11, Physiological Data Collection: Real-time monitor the user's heart rate, respiratory rate, and skin conductance response through wearable devices (smart watches, fitness trackers); S12, Behavioral Data Collection: Record the user's activity trajectory and exercise volume through the GPS unit, accelerometer, and gyroscope built into the mobile device; S13, Psychological State Data Collection: Use a mental health mobile application. The user manually enters their emotional state through the emotion log function, and at the same time, combine with the sleep monitoring unit (sleep stage, wake-up times) of the smart device to automatically collect sleep quality data; S14, Social Interaction Data Collection: By authorizing access to the user's social media accounts, analyze the user's social media activities (posting frequency, number of likes, number of comments), and combine the communication frequency (number of calls, text message frequency) and social network structure (friend relationship) to generate social interaction data.

[0008] Optionally, the data preprocessing in S2 includes: S21, Data Cleaning: Use an outlier detection algorithm based on the interquartile range (IQR) to detect and process outliers in the collected multi-dimensional data; S22, Denoising Processing: Use the wavelet transform method to remove noise in the multi-dimensional data; S23, Missing Value Filling: For the missing values in the multi-dimensional data, use the k-nearest neighbor algorithm for filling; S24, Data Standardization Processing: Use the Z-score standardization method to standardize the multi-dimensional data.

[0009] Optionally, the mental health portrait construction in S3 includes: S31, Multi-dimensional feature extraction: Use the principal component analysis (PCA) method to reduce the dimension of the preprocessed multi-dimensional data and extract multi-dimensional features ; S32, Mental health status scoring: Calculate the mental health score of the user based on the extracted multi-dimensional features ; S33, Risk assessment: Combine multi-dimensional features and predict the mental health risk level of the user through logistic regression ; S34, Intervention need analysis: Generate personalized intervention needs for the user according to the mental health status score and the mental health risk level , specifically including: High-risk users: If and , recommend psychological counseling and professional intervention; Medium-risk users: If and , recommend behavior adjustment and emotion management; Low-risk users: If and , recommend mental health maintenance strategies; S35, Mental health portrait generation: Combine the mental health status score, the mental health score, and the intervention needs to generate a personalized mental health portrait of the user.

[0010] Optionally, the social relationship psychological impact analysis in S4 includes: S41, Social relationship network construction: Based on the user's social interaction data (social media activities, communication frequency, social network structure), construct the user's social relationship network, where network nodes represent the user and their social contact objects, and the edge weights reflect the interaction intensity between the user and adjacent individuals; S42, Emotion propagation and isolation risk analysis: Use the constructed social relationship network to dynamically analyze the emotion propagation path and isolation risk; S43, Generation of psychological impact assessment report: Generate a comprehensive impact assessment report on the mental health of social relationships according to the analysis results of the emotion propagation path and isolation risk.

[0011] Optionally, the social relationship network construction in S41 includes: S411, Data normalization: Extract the key information required to construct the social relationship network from the user's social interaction data, including social media activities (posting frequency, number of likes and comments), communication frequency (number of calls, text message frequency), and social network structure (friend relationship), and perform normalization processing on the social interaction data; S412, Social relationship network construction: Construct an undirected weighted graph through social interaction data , where is a set of nodes, representing users and their social contacts, is a set of edges, representing the social relationships between users and their adjacent individuals, and the edge weights reflect the user and the user The interaction intensity between them; S413, Node feature and edge attribute extraction: In the constructed social relationship network, node features include the user's activity (degree centrality) and influence (betweenness centrality).

[0012] Optionally, the emotion propagation and isolation risk analysis in S42 includes: S421, Emotion propagation path analysis: Using the constructed social relationship network, dynamically analyze the emotion propagation path and key propagation nodes based on the Independent Cascade Model (ICM); S422, Isolation risk analysis: Based on the social relationship network, calculate the user's isolation risk value , when Threshold , identify as a user with high isolation risk.

[0013] Optionally, the generation of the psychological impact assessment report in S43 includes: S431, Quantitative analysis of emotion propagation characteristics: Based on the analysis results of the emotion propagation path, quantify the emotion propagation characteristics and identify the propagation nodes, specifically including: Calculation of propagation coverage rate: Calculate the emotion propagation coverage rate through the total number of activated nodes in the propagation path ; Calculation of propagation intensity: Based on the propagation probability , calculate the propagation intensity of negative emotions ; S432, Comprehensive assessment of isolation risk: According to the results of the isolation risk analysis, calculate the user's network isolation rate, and the network isolation rate is the proportion of all users with high isolation risk within the entire network ; S433, Generation of comprehensive impact assessment report: According to the results of the quantitative analysis of emotion propagation characteristics and the comprehensive assessment of isolation risk, generate a comprehensive impact assessment report on the mental health of social relationships, including propagation coverage rate, propagation intensity, and network isolation rate.

[0014] Optionally, the formulation of the personalized mental health intervention plan in S5 includes: S51, Problem identification and intervention need matching: According to the user's mental health profile and the results of the social relationship psychological impact analysis, identify the user's mental health problems, and directly match the intervention needs in combination with the analysis results, specifically including: Behavior guidance: In view of the high risk of isolation for users, recommend behavior adjustment plans to increase social interaction (participating in offline activities, strengthening connections with specific node users); Emotion management suggestions: For users who spread negative emotions, recommend methods for emotional counseling (meditation, emotion regulation techniques); Social activity recommendations: For users at high risk of isolation, recommend joining interest groups or offline community activities; Psychological counseling suggestions: For users at high risk of isolation, give priority to recommending psychological counseling or psychological intervention; S52, Intervention plan generation: Based on the intervention needs matched by the user, generate a personalized intervention plan, including: Intervention content: Suggestions for each type of intervention (such as meditation training, names of social activities); Execution time and frequency: Provide the schedule and frequency (meditate for 10 minutes every day or participate in an offline activity once a week); Expected effects: Clearly define the intervention goals (improve the mental health score, reduce the risk of isolation).

[0015] Advantages of the present invention: In the present invention, by comprehensively collecting the physiological, behavioral, psychological state and social interaction data of users, combined with advanced data preprocessing methods, the data quality and consistency are greatly improved. Using principal component analysis to reduce the dimension of the data and extract key features, a mental health portrait of the user is constructed and the mental health state and risk level are quantified. It can accurately capture the core influencing factors of the user's mental state, simplify the data dimension while ensuring information integrity, and provide scientific data support for mental health assessment.

[0016] In the present invention, by constructing the user's social relationship network, the emotional transmission path and isolation risk are dynamically analyzed, the transmission intensity and network coverage rate of negative emotions are quantified, and users at high risk of isolation are accurately located. Using the dynamic threshold set by historical data, the mental health risk state of the user in the social relationship is comprehensively evaluated. The generated analysis report on the psychological impact of social relationships not only reveals the potential impact of social interaction on mental health, but also provides a data-driven scientific basis for further personalized mental health intervention.

[0017] In the present invention, by clarifying the intervention content, execution time and frequency, and expected effects, the scientific nature, operability and effectiveness of the intervention plan are ensured. At the same time, the intervention plan supports real-time tracking and dynamic optimization to adapt to the changes in the user's mental state, greatly improving the efficiency and pertinence of mental health management, and providing accurate and efficient mental health intervention services for users. Description of the Drawings

[0018] To more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only those of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0019] Figure 1 Schematic diagram of the evaluation and intervention method for the embodiment of the present invention; Figure 2 Schematic diagram of the analysis of the psychological impact of social relationships for the embodiment of the present invention. Detailed implementation manners

[0020] The present invention will be described in detail below in conjunction with the accompanying drawings and specific embodiments. At the same time, it should be noted here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments. For some well-known technologies, those skilled in the art can also adopt other alternative methods for implementation; moreover, the accompanying drawings are only for more specific description of the embodiments, and are not intended to specifically limit the present invention.

[0021] It should be noted that in the specification, when referring to "an embodiment", "embodiment", "exemplary embodiment", "some embodiments", etc., it indicates that the described embodiment may include specific features, structures or characteristics, but not necessarily every embodiment includes such specific features, structures or characteristics. In addition, when combining embodiments to describe specific features, structures or characteristics, implementing such features, structures or characteristics in combination with other embodiments (whether explicitly described or not) should be within the knowledge of those skilled in the relevant art.

[0022] Generally, the terms can be understood at least in part from their use in the context. For example, at least in part depending on the context, the term "one or more" used herein can be used to describe any feature, structure or characteristic in a singular sense, or can be used to describe a combination of features, structures or characteristics in a plural sense. Additionally, the term "based on" can be understood as not necessarily intended to convey a set of exclusive factors, but rather, at least in part depending on the context, can allow for the existence of other factors that are not necessarily explicitly described.

[0023] As Figure 1 - Figure 2 shown, the AI-based personalized mental health evaluation and intervention method includes the following steps: S1, data collection: Collect multi-dimensional data of users, including physiological data (heart rate, respiratory rate, skin conductance response), behavioral data (activity trajectory, exercise amount), mental state data (emotion log, sleep quality), and social interaction data (social media activities, communication frequency, social network structure); S2, Data preprocessing: Preprocess the collected multi-dimensional data, including cleaning, denoising, missing value imputation, and standardization; S3, Mental health portrait construction: Generate a personalized mental health portrait of the user based on the preprocessed multi-dimensional data, including the user's mental health status score, mental health score, and intervention needs; S4, Analysis of the psychological impact of social relationships: By constructing the user's social relationship network, identify the emotional change patterns, negative emotion propagation paths, and social isolation risks in the user's social interactions, dynamically analyze the emotional propagation characteristics in the social network structure and the potential influencing factors of the user's mental state, and generate an impact assessment report on the influence of social relationships on mental health; S5, Development of personalized mental health intervention programs: Generate personalized mental health intervention programs based on the user's mental health portrait and the results of the analysis of the psychological impact of social relationships, including behavior guidance, emotion management suggestions, social activity recommendations, and psychological counseling suggestions; Through the above content, it is possible to accurately capture the user's mental state and potential influencing factors in social relationships, dynamically adjust the intervention program, ensure the personalization and real-time nature of the intervention, and at the same time make full use of artificial intelligence technology to improve data processing efficiency and evaluation accuracy, providing an efficient, intelligent, and innovative solution for mental health management.

[0024] The data collection in S1 includes: S11, Physiological data collection: Real-time monitor the user's heart rate, respiratory rate, and skin conductance response through wearable devices (smart watches, fitness trackers); S12, Behavioral data collection: Record the user's activity trajectory and exercise volume through the GPS unit, accelerometer, and gyroscope built into the mobile device; S13, Mental state data collection: Use a mental health mobile application. The user manually enters their emotional state through the emotion log function, and at the same time, sleep quality data is automatically collected in combination with the sleep monitoring unit (sleep stage, wake-up times) of the smart device; S14, Social interaction data collection: By authorizing access to the user's social media accounts, analyze the user's social media activities (posting frequency, number of likes, number of comments), and generate social interaction data in combination with the communication frequency (number of calls, text message frequency) and social network structure (friend relationships); Through the above content, it comprehensively covers four key dimensions of physiology, behavior, mental state, and social interaction, ensuring the diversity and accuracy of data sources, being able to capture the user's dynamic state in real time, and providing high-quality multi-dimensional data input.

[0025] The data preprocessing in S2 includes: S21, Data cleaning: An outlier detection algorithm based on the interquartile range (IQR) is used to detect and process outliers in the collected multi-dimensional data, expressed as: Outlier ; Wherein, And Are the first and third quartiles respectively, , Is a single data point or sample value in the multi-dimensional data; S22, Denoising processing: The wavelet transform method is used to remove noise in the multi-dimensional data, expressed as: ; Wherein, Is the signal value at time Moment, Is the total number of layers or the maximum scale number of wavelet decomposition, Is the wavelet decomposition coefficient, representing the contribution of signal At scale And translation position On, Is the wavelet basis function, the wavelet form at scale And translation position On; S23, Missing value filling: For the missing values in the multi-dimensional data, the k-nearest neighbor algorithm is used for filling, expressed as: ; Wherein, Is the value to be filled, Represents the set of the Nearest non-missing values to the missing value, Is the data point in the set; S24, Data standardization processing: The Z-score standardization method is used to standardize the multi-dimensional data, expressed as: ; Wherein, Is the original data value, Is the mean of the data, Is the standard deviation, Is the data value after standardization processing; Through the above content, the quality and consistency of the multi-dimensional data are comprehensively improved. The cleaning step effectively removes outliers, the denoising step retains the key features of the signal and reduces noise interference, the missing value filling ensures the integrity of the data, and the standardization processing eliminates the dimensional differences between different data dimensions.

[0026] The construction of the mental health portrait in S3 includes: S31, multi-dimensional feature extraction: Use the principal component analysis (PCA) method to reduce the dimension of the pre-processed multi-dimensional data and extract multi-dimensional features , specifically including: Covariance matrix calculation: Calculate the covariance matrix of the pre-processed data matrix , expressed as: ; ; where, is the number of samples, is the transpose of the data matrix; Principal component extraction: Perform eigenvalue decomposition on the covariance matrix to obtain eigenvalues and the corresponding eigenvectors, and select the first eigenvalues whose cumulative contribution rate reaches 95% and their corresponding vectors, expressed as: Cumulative contribution rate ; where, is the th eigenvalue, is the total number of original data features; Feature generation: Project the data matrix onto the reduced-dimensional eigenvectors to generate multi-dimensional features , expressed as: ; where, is the th eigenvector; S32, mental health status scoring: Based on the extracted multi-dimensional features, calculate the mental health score of the user , expressed as: ; where, is the weight of the multi-dimensional feature , is the number of multi-dimensional features; S33, risk assessment: Combine multi-dimensional features and predict the mental health risk level of the user through logistic regression , expressed as: ; where, is the logistic function, is the model weight, is the bias term, represents the probability value of the user's mental health risk; S34, Intervention Requirement Analysis: Generate the personalized intervention requirements of the user based on the mental health status score and the mental health risk level , specifically including: High-risk users: If and , recommend psychological counseling and professional intervention; Medium-risk users: If and , recommend behavior adjustment and emotion management; Low-risk users: If and , recommend mental health maintenance strategies; S35, Mental Health Portrait Generation: Generate the personalized mental health portrait of the user by combining the mental health status score, the mental health score, and the intervention requirements Through the above content, the core influencing factors of the user's mental state can be accurately captured, the data dimension can be simplified while retaining key information, and scientific and quantitative health scores and risk analysis results can be provided. Through the dynamic update of the portrait and the generation of personalized intervention requirements, the accuracy, real-time performance, and practicality of mental health assessment and intervention are significantly improved, providing reliable data support and decision-making basis for individualized health management.

[0027] The psychological impact analysis in S4 includes: S41, Social Relationship Network Construction: Based on the user's social interaction data (social media activities, communication frequency, social network structure), construct the user's social relationship network. The network nodes represent the user and their social contact objects, and the edge weights reflect the interaction intensity between the user and adjacent individuals S42, Emotion Propagation and Isolation Risk Analysis: Use the constructed social relationship network to dynamically analyze the emotion propagation path and isolation risk S43, Generation of Psychological Impact Assessment Report: Generate a comprehensive impact assessment report on the mental health of social relationships based on the analysis results of the emotion propagation path and isolation risk Through the above content, the accurate analysis of the psychological impact in the user's social interaction is realized. It can identify key propagation nodes, the propagation trend of negative emotions, and high-isolation-risk users, thus comprehensively revealing the potential impact of social interaction on mental health. Combining the user's social data and dynamic mental state characteristics, it provides quantitative and personalized mental health assessments, providing a scientific basis for precise intervention and improvement of social interaction, and significantly enhancing the efficiency and effectiveness of mental health management.

[0028] The social relationship network construction in S41 includes: S411, Data Normalization: Extract the key information required to construct a social relationship network from the user's social interaction data, including social media activities (posting frequency, number of likes and comments), communication frequency (number of calls, text message frequency), and social network structure (friend relationships), and perform normalization processing on the social interaction data, expressed as: ; Among them, is the normalized social interaction data, is the original social interaction data, and are the minimum and maximum values of the social interaction data respectively; S412, Social Relationship Network Construction: Construct an undirected weighted graph , where is the node set, representing the user and their social contact objects, is the edge set, representing the social relationships between the user and their adjacent individuals, and the edge weight reflects the interaction intensity between user and user , expressed as: ; Among them, represents the social media interaction frequency between user and user , represents the communication frequency between user and user , represents the structural relationship between user and user in the social network, , , are weight coefficients; S413, Node Feature and Edge Attribute Extraction: In the constructed social relationship network, node features include the user's activity (degree centrality) and influence (betweenness centrality), expressed as: ; Among them, is the neighbor node set of user , is the degree centrality of user , representing the activity level of user in the social relationship network, is the total connection strength of user ; ; Among them, Indicates the number of shortest paths from node to node . Indicates the number of shortest paths passing through node . Is the betweenness centrality of user , indicating the degree to which user serves as a bridge between other users in the social relationship network; Through the above content, the social interaction data of users is integrated and the interaction intensity between users is dynamically quantified using edge weights, comprehensively depicting the social relationship characteristics of users. Based on the degree centrality and betweenness centrality analysis in the network, it is not only possible to quantify the activity level and global influence of users, but also to reveal the key nodes and potential isolated individuals in the social relationship.

[0029] The emotion propagation and isolation risk analysis in S42 includes: S421, emotion propagation path analysis: Using the constructed social relationship network, dynamically analyze the emotion propagation path and key propagation nodes based on the Independent Cascade Model (ICM), expressed as: ; Among them, Indicates the probability that the emotion of node propagates to node , Is the edge weight between node and node , Is the set of neighbor nodes of node , Is the edge weight between node and node ; The propagation process simulation includes: (1) At the initial moment, emotion propagation starts from a group of activated nodes (negative emotion sources); (2) Each activated node tries to activate its neighbor node with probability ; (3) If the activation is successful, node becomes part of the propagation path and continues to propagate. If the activation fails, the propagation terminates; S422, isolation risk analysis: Based on the social relationship network, calculate the isolation risk value of the user. When is greater than the threshold , the user is identified as a high isolation risk user, expressed as: ; Among them, Represents the isolation risk value of the user , and is the degree centrality of the user ; is the maximum degree centrality in the network; The threshold is set based on historical data, specifically including: Collect historical data: Extract isolation risk indicators and mental health status data related to user mental health from a large amount of historical social network data; Isolation risk indicator calculation: Calculate the isolation risk values of all users and normalize the isolation risk values; Analyze the historical data distribution: Divide the isolation risk values of all users into three groups (high risk, medium risk, low risk) according to their actual mental health status, and analyze the statistical distribution characteristics (mean and standard deviation) of the isolation risk values in each group, expressed as: ; ; ; ; Among them, and are the mean and standard deviation of high-risk users respectively, and are the mean and standard deviation of low-risk users respectively, and are the numbers of users in the high-risk group and the low-risk group respectively; Threshold setting: Use the distribution characteristics of the isolation risk values of the high-risk group to set the threshold, expressed as: ; Among them, is the adjustment coefficient (taking values from 1 to 2); Through the above content, not only can the propagation nodes and key influential individuals of negative emotions be identified, but also high-isolation-risk users can be accurately located. By combining historical data to set dynamic thresholds, a comprehensive quantitative assessment of the user's mental health status is provided, realizing in-depth mining of complex social interactions and psychological impacts, providing precise support for personalized mental health interventions, and significantly improving the timeliness of analysis and the scientificity of decision-making.

[0030] The generation of the psychological impact assessment report in S43 includes: S431, Quantitative analysis of emotion propagation characteristics: Based on the analysis results of the emotion propagation path, quantify the emotion propagation characteristics and identify the propagation nodes, specifically including: Calculation of the propagation coverage rate: Calculate the emotional propagation coverage rate through the total number of activated nodes in the propagation path , which is expressed as: ; Among them, represents the total number of nodes activated by the propagated emotion, represents the total number of nodes in the social network; Calculation of the propagation intensity: Based on the propagation probability , calculate the propagation intensity of negative emotions , which is expressed as: ; Among them, is the emotional propagation probability that node propagates to node , represents the negative emotion weight from node to node ; S432, Comprehensive assessment of isolation risk: According to the results of the isolation risk analysis, calculate the user's network isolation rate, which is the proportion of all users with high isolation risk in the entire network , which is expressed as: ; Among them, represents the number of users with an isolation risk value higher than the threshold, represents the total number of users in the social network; S433, Generation of a comprehensive impact assessment report: According to the results of the quantitative analysis of emotional propagation characteristics and the comprehensive assessment of isolation risk, generate a comprehensive impact assessment report on the mental health of social relationships, including the propagation coverage rate, propagation intensity, and network isolation rate; Through the above content, the potential impact of social relationships on users' mental health is comprehensively quantified. Using the propagation coverage rate, propagation intensity, and isolation risk indicators, not only can the propagation path and key nodes of negative emotions be accurately located, but also users with high isolation risk can be identified, revealing the weak links in their mental health status. The generated assessment report provides clear quantitative results and personalized intervention suggestions, realizing multi-dimensional analysis, quantification, and visualization of psychological impacts, providing a scientific basis and precise support for mental health management, and greatly improving the efficiency and pertinence of mental health intervention.

[0031] The formulation of the personalized mental health intervention plan in S5 includes: S51, Matching problem identification with intervention needs: According to the user's mental health portrait and the results of the psychological impact analysis of social relationships, identify the user's mental health problems, and directly match the intervention needs in combination with the analysis results. Specifically, it includes: Behavior guidance: For users at high risk of isolation, recommend behavior adjustment plans to increase social interaction (participating in offline activities, strengthening connections with specific node users); Emotion management suggestions: For users who spread negative emotions, recommend emotion counseling methods (meditation, emotion regulation techniques); Social activity recommendations: For users at high risk of isolation, recommend joining interest groups or offline community activities; Psychological counseling suggestions: For users at high risk of isolation, prioritize psychological counseling or psychological intervention; S52, Intervention plan generation: Based on the intervention needs matched by the user, generate a personalized intervention plan, including: Intervention content: Suggestions for each type of intervention (such as meditation training, social activity names); Execution time and frequency: Provide the schedule and frequency (meditate for 10 minutes every day or participate in an offline activity once a week); Expected effects: Clearly define the intervention goals (improve mental health score, reduce the risk of isolation); Based on the above content, combined with the mental health portrait and the analysis results of the psychological impact of social relationships, generate personalized intervention needs including behavior guidance, emotion management, social activity recommendations, and psychological counseling suggestions. At the same time, through the formulation of a detailed intervention plan, clarify the intervention content, execution time and frequency, and expected effects to ensure the scientific nature and operability of the intervention.

[0032] This invention covers any substitutions, modifications, equivalent methods, and solutions made within the essence and scope of this invention. To enable the public to have a thorough understanding of this invention, specific details are described in detail in the following preferred embodiments of this invention. However, those skilled in the art can fully understand this invention without the description of these details. Additionally, to avoid unnecessary confusion to the essence of this invention, well-known methods, processes, procedures, components, and circuits are not described in detail.

[0033] The above are only the preferred embodiments of this invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of this invention, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of this invention.

Claims

1. AI-based personalized mental health assessment and intervention method, characterized by: The following steps are involved: S1, data collection: collect multi-dimensional data of users, including physiological data, behavioral data, psychological state data and social interaction data; S2, data preprocessing: preprocess the collected multi-dimensional data, including cleaning, denoising, missing value filling and standardization; S3, mental health portrait construction: Based on the pre-processed multi-dimensional data, generate a personalized mental health portrait of the user, including the user's mental health status score, mental health score and intervention needs, including: S31, multi-dimensional feature extraction: Use the principal component analysis method to reduce the dimension of the pre-processed multi-dimensional data and extract multi-dimensional features ; S32, mental health status score: Calculate the user's mental health score based on the extracted multi-dimensional features ; S33, Risk Assessment: Combining multi-dimensional features, predicting the user's mental health risk level through logistic regression ; S34, Intervention Needs Analysis: Generate personalized intervention needs for users based on mental health status scores and mental health risk levels , specifically including: High-risk users: If and , recommend psychological counseling and professional intervention; Medium risk users: If and , recommend behavioral adjustments and emotion management; Low-risk users: If and , recommend strategies for maintaining mental health; S35, mental health portrait generation: Generate a personalized mental health portrait of the user by combining the mental health status score, mental health score and intervention needs; S4, Analysis of the psychological impact of social relationships: By building a user's social relationship network, identifying the user's emotional change patterns, negative emotional transmission paths, and social isolation risks in social interactions, dynamically analyzing the emotional transmission characteristics in the social network structure and the potential influencing factors of the user's psychological state, and generating an assessment report on the impact of social relationships on mental health; S5, formulation of personalized mental health intervention plan: based on the user's mental health portrait and the results of social relationship psychological impact analysis, generate a personalized mental health intervention plan, including behavioral guidance, emotion management suggestions, social activity recommendations, and psychological counseling suggestions.

2. The AI-based personalized mental health assessment and intervention method according to claim 1, characterized in that: The data collection in S1 includes: S11, physiological data collection: real-time monitoring of the user's heart rate, breathing rate and skin galvanic response through wearable devices; S12, behavioral data collection: recording the user's activity trajectory and exercise volume through the built-in GPS unit, accelerometer and gyroscope of the mobile device; S13, psychological state data collection: using the psychological health mobile application, users manually input their emotional state through the emotional log function, and the sleep quality data is automatically collected in conjunction with the sleep monitoring unit of the smart device; S14, social interaction data collection: by authorizing access to the user's social media account, analyzing the user's social media activities, and generating social interaction data based on communication frequency and social network structure.

3. The AI-based personalized mental health assessment and intervention method according to claim 1, characterized in that: The data preprocessing in S2 includes: S21, data cleaning: using an outlier detection algorithm based on the quartile range to detect and process outliers in the collected multi-dimensional data; S22, denoising: removing noise from multi-dimensional data using wavelet transform method; S23, missing value filling: For missing values ​​in multi-dimensional data, the k-nearest neighbor algorithm is used to fill them; S24, data standardization: use Z-score standardization method to standardize multi-dimensional data.

4. The AI-based personalized mental health assessment and intervention method according to claim 1, characterized in that: The social relationship psychological impact analysis in S4 includes: S41, social relationship network construction: Based on the user's social interaction data, the user's social relationship network is constructed. The network nodes represent the user and his social contact objects, and the edge weights reflect the interaction intensity between the user and the adjacent individuals; S42, Emotional propagation and isolation risk analysis: Using the constructed social relationship network, dynamically analyze the emotion propagation path and isolation risk; S43, generation of psychological impact assessment report: Based on the analysis results of emotion propagation paths and isolation risks, a comprehensive impact assessment report of social relationships on mental health is generated.

5. The AI-based personalized mental health assessment and intervention method according to claim 4, characterized in that: The social relationship network construction in S41 includes: S411, data normalization: extract key information needed to build a social relationship network from users' social interaction data, including social media activities, communication frequency, and social network structure, and normalize the social interaction data; S412, Social Relationship Network Construction: Constructing Undirected Weighted Graphs from Social Interaction Data ,in is a node set, representing users and their social connection objects, is an edge set, representing the social relationship between the user and his adjacent individuals, and the edge weight Reflecting users With users The intensity of interaction between them; S413, Node feature and edge attribute extraction: In the constructed social relationship network, node features include user activity and influence.

6. The AI-based personalized mental health assessment and intervention method according to claim 5, characterized in that: The emotion propagation and isolation risk analysis in S42 includes: S421, Emotional Propagation Path Analysis: Using the constructed social relationship network, the emotional propagation path and key propagation nodes are dynamically analyzed based on the independent cascade model; S422, Isolation Risk Analysis: Calculate the user's isolation risk value based on the social relationship network ,when Threshold When a user is identified as a high risk user.

7. The AI-based personalized mental health assessment and intervention method according to claim 6, characterized in that: The generation of the psychological impact assessment report in S43 includes: S431, quantitative analysis of emotion propagation characteristics: based on the analysis results of emotion propagation paths, quantify emotion propagation characteristics and identify propagation nodes, including: Propagation coverage calculation: The sentiment propagation coverage is calculated by the total number of activated nodes in the propagation path. ; Propagation intensity calculation: based on propagation probability , calculate the propagation intensity of negative emotions ; S432, Comprehensive evaluation of isolation risk: Based on the results of isolation risk analysis, the user's network isolation rate is calculated. The network isolation rate is the proportion of all high-isolation risk users in the entire network. ; S433, generation of comprehensive impact assessment report: Based on the results of quantitative analysis of emotion propagation characteristics and comprehensive assessment of isolation risk, generate a comprehensive impact assessment report on social relationships on mental health, including propagation coverage, propagation intensity, and network isolation rate.

8. The AI-based personalized mental health assessment and intervention method according to claim 7, characterized in that: The development of a personalized mental health intervention program in S5 includes: S51, problem identification and intervention needs matching: Based on the user's mental health portrait and the results of the social relationship psychological impact analysis, identify the user's mental health problems, and directly match the intervention needs based on the analysis results, including: Behavioral guidance: For users with high isolation risk, behavioral adjustment plans to increase social interaction are recommended; Emotion management suggestions: Recommend emotional guidance methods for users who spread negative emotions; Social activity recommendations: For users at high risk of isolation, we recommend them to join interest groups or offline community activities; Psychological counseling suggestions: For users with high isolation risk, psychological counseling or psychological intervention is recommended; S52, Intervention Plan Generation: Generate a personalized intervention plan based on the intervention needs matched by the user, including: Intervention content: recommendations for each intervention type; Execution time and frequency: Provide the time schedule and frequency; Desired outcome: Clarify intervention objectives.

Images