Mental health intervention method and device, electronic equipment and storage medium

Abstract

The invention provides a mental health intervention method and device, electronic equipment and a storage medium. The method provided by the invention comprises the steps of obtaining evaluation behavior characteristic data of a user, the evaluation behavior characteristic data comprising an answer consistency score, a completion time rationality score and a response mode regularity score; determining a credibility index according to the evaluation behavior feature data, determining an early warning level according to the credibility index, and determining a preset first intervention strategy according to the early warning level; after a preset first intervention strategy acts on the user, obtaining user feature data, and forming a user feature vector according to the user feature data; and through the trained classification model and the clustering category of the user, allocating a corresponding preset second intervention strategy to the user to perform psychological intervention and obtain an intervention effect label, and if the level of the intervention effect label is lower than a preset label level, stopping pushing the preset second intervention strategy to the user. According to the method, the mental health intervention effect can be improved.

Description

Technical Field

[0001] The present invention relates to the field of mental health technology, and in particular to a mental health intervention method, device, electronic device and storage medium. Background Art

[0002] Traditional mental health assessments rely primarily on face-to-face diagnoses and paper-and-pencil questionnaires. These methods are not only time-consuming and costly, but also susceptible to subjective factors, making it difficult to monitor changes in a patient's mental state in real time. Existing mental health systems primarily construct user profiles based on psychological assessment results, lacking in-depth mining and analysis of multidimensional user behavioral data. They rely solely on current mental state for early warning and assessment, resulting in poor intervention effectiveness. Summary of the Invention

[0003] The purpose of the present invention is to provide a mental health intervention method, device, electronic device and storage medium to solve the technical problem of poor effect of mental health intervention in the prior art.

[0004] The technical solution of the present invention is as follows, which provides a mental health intervention method, comprising: Obtaining user assessment behavior characteristic data, wherein the assessment behavior characteristic data includes a consistency score for answering questions, a reasonableness score for completion time, and a regularity score for response patterns; Determining a credibility index based on the assessed behavioral characteristic data, determining a warning level based on the credibility index, and determining a preset first intervention strategy based on the warning level; After the preset first intervention strategy acts on the user, obtaining user feature data, and forming a user characteristic vector according to the user feature data; The user's clustering category is determined based on the user's characteristic vector. Through the trained classification model and the user's clustering category, the corresponding preset second intervention strategy is assigned to the user to perform psychological intervention and obtain an intervention effect label. If the level of the intervention effect label is lower than the preset label level, the preset second intervention strategy is stopped from being pushed to the user; the classification model is trained based on the user's characteristic vector sample and the preset second intervention strategy.

[0005] Furthermore, determining a credibility index based on the evaluation behavior characteristic data, and determining a warning level based on the credibility index, includes: A weighted sum is performed on the answer consistency score, the completion time rationality score and the response pattern regularity score to obtain a credibility index. If the credibility index is not greater than the first credibility threshold, the warning level is determined to be the first level; if the credibility index is greater than the first credibility threshold and not greater than the second credibility threshold, the warning level is determined to be the second level; if the credibility index is greater than the second credibility threshold and not greater than the third credibility threshold, the warning level is determined to be the third level; if the credibility index is greater than the third credibility threshold and not greater than the fourth credibility threshold, the warning level is determined to be the fourth level; if the credibility index is greater than the fourth credibility threshold, the warning level is determined to be the fifth level.

[0006] Furthermore, a preset first intervention strategy is determined according to the warning level, including: If the warning level is level one, the preset first intervention strategy is regular monitoring; if the warning level is level two, the preset first intervention strategy is the use of a self-help resource package; if the warning level is level three, the preset first intervention strategy is the use of a combination of artificial intelligence consultation and self-help tools; if the warning level is level four, the preset first intervention strategy is the use of professional psychological counselors and auxiliary tools; if the warning level is level five, the preset first intervention strategy is emergency medical referral and immediate psychological intervention.

[0007] Furthermore, obtaining user feature data and forming a user characteristic vector based on the user feature data includes: Obtain course learning and interactive behavior dimension data, self-service tool usage and interaction dimension data, artificial intelligence consultation multimodal interaction dimension data, and professional consultation feedback integration dimension data; use at least one dimension data among the course learning and interactive behavior dimension data, the self-service tool usage and interaction dimension data, the artificial intelligence consultation multimodal interaction dimension data, and the professional consultation feedback integration dimension data as user feature data, and form a user feature vector based on the user feature data.

[0008] Furthermore, the step of training the classification model includes: Clustering the user feature vector samples to determine the cluster categories of the corresponding users, using the same preset second intervention strategy to perform psychological intervention on users in the same cluster category, obtaining intervention effect labels, and optimizing the intervention effect labels by adjusting the preset second intervention strategy.

[0009] Furthermore, using the same preset second intervention strategy to perform psychological intervention on users in the same cluster category to obtain intervention effect labels includes: Utilize the same preset second intervention strategy to conduct psychological intervention on users in the same cluster category, track the immediate effect and short-term effect of the intervention, obtain the immediate effect score and short-term effect score respectively, determine the comprehensive effect score based on the immediate effect score and the short-term effect score, and obtain the intervention effect label based on the comprehensive effect score.

[0010] Furthermore, the immediate effect tracking includes monitoring emotional changes, platform engagement, behavioral responses and subjective satisfaction, and the short-term effect tracking includes monitoring symptom improvement effects, functional recovery effects, treatment compliance and side effects.

[0011] Another technical solution of the present invention is as follows: a mental health intervention device is provided, comprising a data acquisition module, a warning level determination module, a characteristic vector generation module, and an intervention module; The data acquisition module is used to obtain the user's evaluation behavior characteristic data, wherein the evaluation behavior characteristic data includes the answer consistency score, the completion time rationality score and the response pattern regularity score; The warning level determination module is configured to determine a credibility index based on the assessed behavior characteristic data, determine a warning level based on the credibility index, and determine a preset first intervention strategy based on the warning level; The characteristic vector generating module is configured to obtain user characteristic data after the preset first intervention strategy acts on the user, and form a user characteristic vector based on the user characteristic data; The intervention module is used to determine the user's cluster category based on the user characteristic vector, and assign the corresponding preset second intervention strategy to the user through the trained classification model and the user's cluster category to perform psychological intervention and obtain an intervention effect label. If the level of the intervention effect label is lower than the preset label level, the preset second intervention strategy is stopped from being pushed to the user; the classification model is trained based on the user characteristic vector sample and the preset second intervention strategy.

[0012] Another technical solution of the present invention is as follows: an electronic device is provided, including a memory and a processor, wherein the memory stores a computer program that can be executed by the processor, and when the processor executes the computer program, the mental health intervention method described in any of the above technical solutions is implemented.

[0013] Another technical solution of the present invention is as follows: a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the mental health intervention method as described in any of the above technical solutions is implemented.

[0014] The beneficial effects of the present invention are: obtaining the user's evaluation behavior characteristic data, the evaluation behavior characteristic data including the answer consistency score, the completion time rationality score and the response pattern regularity score; determining the credibility index according to the evaluation behavior characteristic data, determining the warning level according to the credibility index, and determining the preset first intervention strategy according to the warning level; after the preset first intervention strategy acts on the user, obtaining the user characteristic data, and forming a user characteristic vector according to the user characteristic data; determining the user's clustering category according to the user characteristic vector, and assigning the corresponding preset second intervention strategy to the user through the trained classification model and the user's clustering category to perform psychological intervention and obtain the intervention effect label. If the level of the intervention effect label is lower than the preset label level, then stop pushing the preset second intervention strategy to the user; the classification model is trained based on the user characteristic vector sample and the preset second intervention strategy; through the above technical solution, early warning and intervention of mental health can be achieved, the rationality of mental health intervention can be improved, and thus the effect of mental health intervention can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 A flowchart of a mental health intervention method provided in an embodiment of the present invention.

[0016] Figure 2 A schematic diagram of the structure of a mental health intervention device provided in an embodiment of the present invention.

[0017] Figure 3 A schematic structural diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0018] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0019] References herein to "embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present invention. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute a separate or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0020] Figure 11 is a flow chart of a mental health intervention method according to an embodiment of the present invention. It should be noted that the mental health intervention method of the present invention is not limited to the method of the embodiment of the present invention if the results are substantially the same. Figure 1 The process sequence shown is limited. Figure 1 As shown, the mental health intervention method mainly includes the following steps: S101, obtaining user's assessment behavior characteristic data, wherein the assessment behavior characteristic data includes answer consistency score, completion time rationality score and response pattern regularity score; In a specific embodiment, standardized psychological scale assessment and scale combination assessment are implemented on the user, and the user's assessment behavior characteristic data is recorded in real time. The assessment behavior characteristic data may include answer consistency score, completion time rationality score and response pattern regularity score.

[0021] As an implementation method, the answer consistency score can be calculated as follows: For example, question A is "I often feel anxious" (positive scoring, 0-3 points), and question B is "I rarely worry about the future" (reverse scoring, 0-3 points); 0 is never / not at all, 1 is rarely / mild, 2 is often / moderate, and 3 is always / severe; assuming the user answers, question A is 2 points (often feeling anxious) and question B is 1 point (rarely worrying about the future), the positive score = option score, the reverse score = (highest score + lowest score) - option score, the reverse score is the converted score of question B, the maximum possible difference = interval normalization, the answer consistency score = 1-(2-2) / 3=1, and the answer consistency score of 1 indicates complete consistency.

[0022] As an implementation method, the reasonableness of a scale's completion time refers to whether the total time a participant spends completing a self-assessment scale is consistent with the number of questions, the complexity of the questions, and the respondent's average reading and comprehension speed. A reasonable completion time reflects the participant's thoroughness in answering the questions and the credibility of the data, making it a crucial component of quality control. Each self-assessment scale contains different questions; the more questions, the longer the completion time. Some scale questions are short, such as "I feel nervous," while others are longer, such as "In the past two weeks, have I often been upset by trivial matters?" Longer questions require more time to read and comprehend. Assuming a diligent respondent, the average reading and reaction time per question is approximately 5-10 seconds. This value may vary depending on the respondent's age, educational background, and question complexity. The completion time reasonableness score (for determining abnormal responses) can be calculated as follows: Completion time reasonableness score = Overall time reasonableness calculation weight × Time reasonableness score + Time distribution uniformity calculation weight × Time distribution score + Extreme time penalty weight × Extreme time penalty score. As an example, the weight for calculating the rationality of the overall time is 0.5, the minimum and maximum times of the overall time are set, the reasonable range is [minimum time, maximum time], the weight for calculating the uniformity of time distribution is 0.3, the time distribution score = 1 / (1+answer variance / historical variance), the extreme time penalty weight is 0.2, the standard time = 9×20 = 180 seconds, the actual time = 138 seconds, which is within the reasonable range [90, 540], the time rationality score = 1.0, the time variance = var([12,8,15,20,10,18,14,25,16]) = 25.8, the reference variance = 30 (based on historical data), the time distribution score = 1 / (1+25.8 / 30) = 0.54, there is no extreme time (if there is no extreme time), the extreme time penalty score = 1.0, and the completion time rationality score = 0.5×1.0 + 0.3×0.54 +0.2×1.0 = 0.862.

[0023] As an implementation method, the regularity of the response pattern is used to determine whether the user has a logical answer trajectory in the process of answering questions, reflecting whether the user's answering behavior is regular and stable, and whether there are obvious abnormal response behaviors (such as random clicking, repeatedly selecting the same option, etc.). The response pattern regularity score can be determined by calculating the entropy. The larger the entropy, the more diverse. The larger the entropy value, the greater the uncertainty, the more uniform the distribution, and the better the diversity; the smaller the entropy value, the greater the certainty, the more concentrated the distribution, and the worse the diversity. As an example, the total number of answers = 9 questions, the option distribution: 1 (0 times), 2 (2 times), 3 (5 times), 4 (2 times), 5 (0 times), the probability of each option, = 0 / 9 = 0, = 2 / 9 ≈ 0.22, = 5 / 9 ≈0.56, = 2 / 9 ≈ 0.22, = 0 / 9 = 0, answer entropy (actual entropy): H = -[0×0 + 0.22×(-1.50)+ 0.56×(-0.65) + 0.22×(-1.50) + 0×0] = -[0 + (-0.33) + (-0.364) + (-0.33)+ 0]= -(-1.024) = 1.024 ≈ 1.027, when all options are equally likely to be selected, the entropy reaches its maximum value. For 5 options, the probability of each option = 1 / 5 = 0.2, the maximum entropy = -5 × (0.2 × log(0.2)) = -5 × (0.2× (-1.609)) = -5 × (-0.322) = 1.609. Response pattern regularity score (diversity score) = actual entropy / maximum entropy = 1.027 / 1.609 = 0.64. Diversity score rating: 0.8-1.0 corresponds to excellent diversity, and answers are very serious; 0.6-0.8 corresponds to good diversity, with some preferences but reasonable; 0.4-0.6 corresponds to average diversity, with obvious preferences; 0.2-0.4 corresponds to poor diversity, and there may be response bias; 0.0-0.2 corresponds to very poor diversity, and there may be random answers or extreme response tendencies.

[0024] S102, determining a credibility index based on the assessed behavioral characteristic data, determining a warning level based on the credibility index, and determining a preset first intervention strategy based on the warning level; In some embodiments, determining a credibility index based on the assessed behavioral characteristic data, and determining a warning level based on the credibility index includes: A weighted sum is performed on the answer consistency score, the completion time rationality score and the response pattern regularity score to obtain a credibility index. If the credibility index is not greater than the first credibility threshold, the warning level is determined to be the first level; if the credibility index is greater than the first credibility threshold and not greater than the second credibility threshold, the warning level is determined to be the second level; if the credibility index is greater than the second credibility threshold and not greater than the third credibility threshold, the warning level is determined to be the third level; if the credibility index is greater than the third credibility threshold and not greater than the fourth credibility threshold, the warning level is determined to be the fourth level; if the credibility index is greater than the fourth credibility threshold, the warning level is determined to be the fifth level.

[0025] In a specific embodiment, the credibility index = ×Answer consistency score+ ×Completion time reasonableness score+ ×Response pattern regularity score, e.g. = 0.4, = 0.3, = 0.3 (weight coefficient). Different warning levels can be displayed in different colors on the user interface, with dynamic grading thresholds set. Level 1 (corresponding to dark green) indicates excellent psychological state and no special intervention is required. Level 2 (corresponding to light green) indicates mild psychological stress and can be recommended for self-help resources. Level 3 (corresponding to yellow) indicates moderate psychological risk and can be assisted by AI consultation and intervention. Level 4 (corresponding to orange) indicates high psychological risk and can be treated with professional manual consultation. Level 5 (corresponding to red) indicates severe psychological crisis and requires urgent medical referral.

[0026] In some embodiments, determining a preset first intervention strategy based on the warning level includes: If the warning level is level one, the preset first intervention strategy is regular monitoring; if the warning level is level two, the preset first intervention strategy is the use of a self-help resource package; if the warning level is level three, the preset first intervention strategy is the use of a combination of artificial intelligence consultation and self-help tools; if the warning level is level four, the preset first intervention strategy is the use of professional psychological counselors and auxiliary tools; if the warning level is level five, the preset first intervention strategy is emergency medical referral and immediate psychological intervention.

[0027] In a specific embodiment, a preset first intervention strategy is determined based on the warning level. If the warning level is level one, the preset first intervention strategy is regular monitoring, the monitoring frequency may be once every two weeks, and the method may be a lightweight assessment. If the warning level is level two, the preset first intervention strategy is the use of a self-help resource package, which may include a psychological course library, meditation exercises, reading materials, and emotion management tools. If the warning level is level three, the preset first intervention strategy is the use of an artificial intelligence consultation and a self-help tool combination, which may include in-depth conversations with chatbots and personalized exercises. If the warning level is level four, the preset first intervention strategy is the use of professional psychological counselors and auxiliary tools, which may include AI-assisted analysis. If the warning level is level five, the preset first intervention strategy is emergency medical referral and immediate psychological intervention, which may include a hospital green channel and 24-hour monitoring.

[0028] S103, after the preset first intervention strategy acts on the user, obtaining user feature data, and forming a user feature vector according to the user feature data; In some embodiments, obtaining user feature data and forming a user characteristic vector based on the user feature data includes: Obtain course learning and interactive behavior dimension data, self-service tool usage and interaction dimension data, artificial intelligence consulting multimodal interaction dimension data, and professional consulting feedback integration dimension data, use at least one of the course learning and interactive behavior dimension data, the self-service tool usage and interaction dimension data, the artificial intelligence consulting multimodal interaction dimension data, and the professional consulting feedback integration dimension data as user feature data, and form a user feature vector based on the user feature data.

[0029] In a specific embodiment, the dimensions of course learning and interactive behavior may include learning engagement, interactive activity, knowledge absorption effect, and content preference; the dimensions of self-service tool usage and interaction may include tool usage time, category, content, and behavioral data; the dimensions of AI (artificial intelligence) consultation multimodal interaction may include text sentiment evolution, topic distribution, help content recognition, voice emotion, emotion consistency, and multimodal features; the dimensions of professional consultation feedback integration may include professional profile and counselor confidence, where professional profile = [diagnostic label, treatment recommendation, prognosis assessment, risk level, intervention focus] and counselor confidence = [assessment certainty, treatment response expectation, recurrence risk assessment]. User feature data under the above dimensions can all be quantified using values ​​of [0,1].

[0030] S104, determining the user's clustering category based on the user characteristic vector, and assigning the corresponding preset second intervention strategy to the user through the trained classification model and the user's clustering category to perform psychological intervention and obtain an intervention effect label. If the level of the intervention effect label is lower than the preset label level, stopping pushing the preset second intervention strategy to the user; the classification model is trained based on the user characteristic vector sample and the preset second intervention strategy.

[0031] In some embodiments, the step of training the classification model includes: Clustering the user feature vector samples to determine the cluster categories of the corresponding users, using the same preset second intervention strategy to perform psychological intervention on users in the same cluster category, obtaining intervention effect labels, and optimizing the intervention effect labels by adjusting the preset second intervention strategy.

[0032] In one specific embodiment, K-means clustering is used to achieve preliminary grouping of user groups and pre-assignment of strategies. Then, by tracking intervention effects, labels are generated, ultimately training a classification model that can predict the matching effect between users and corresponding strategies, achieving intervention decision optimization from experience-driven to data-driven. Specifically, the user feature vector samples can be clustered to obtain the corresponding user cluster category. The user feature vector sample can be X = [learning engagement, interactive activity, knowledge absorption effect, content preference, tool usage time, tool category preference, text sentiment mean, topic distribution, voice sentiment consistency, professional diagnosis label, consultant confidence, ...]. The minimization objective function of K-means clustering is , is the i-th user feature vector sample, is the jth category, is the weight corresponding to the i-th feature and the j-th category. The same preset second intervention strategy is used to perform psychological intervention on users in the same cluster category to obtain an intervention effect label. An initial preset second intervention strategy can be set, and the preset second intervention strategy can be adjusted to optimize the intervention effect label.

[0033] In a specific embodiment, after clustering users using the K-means clustering algorithm, each clustering result can be further semanticized to form a number of cluster labels (clustering categories) with guiding significance for psychological intervention. These cluster labels include: high involvement-low distress type (CL-A), this type of users have high learning investment, relatively stable emotional state, low psychological assessment risk level, and active platform interaction. It is recommended to use self-service growth resource packages and advanced psychology courses as intervention methods; high distress-active help-seeking type (CL-B), this type of users show medium to high psychological risk levels and can actively use AI consultation. The emotional density in their conversation texts is high. It is recommended to use a combination of AI consultation and self-service tools and incorporate them into a continuous intervention process; low activity-high defensive type (CL-C), this type of users use tools less frequently, have less interactive behavior, limited emotional expression, and often show defensive or avoidant attitudes. The intervention strategy should focus on building a sense of security and trust, supplemented by lightweight interactive task guidance; emotional sensitivity-low learning investment type (CL-D), this type of users have obvious emotional fluctuations, but low learning participation in the platform, short tool usage time, and prefer to focus on negative emotional topics. Emotional stabilization training and short-term effectiveness exercises are suitable for this type of users; high anxiety-logical self-consistency type (CL-E), this type of users have high psychological test scores High: users with relatively regular answering patterns and positive feedback often exhibit high anxiety but possess good self-reflection abilities. Cognitive reconstruction intervention modules can be prioritized, and the effectiveness of AI dialogue feedback can be enhanced. High-Frequency Tool Type (CL-F): users frequently use self-help tools, but their use lacks a clear purpose and often involves repetitive operations. Intervention requires optimizing the tool combination structure and enhancing their self-awareness through goal guidance. Multimodal Conflict Type (CL-G): users of this type appear low risk in the scale assessment, but their multimodal data, such as AI dialogue or voice, show a high degree of repression or conflict, indicating potential psychological distress. It is recommended that multimodal emotion recognition results be combined with manual professional judgment to provide intervention support. Professional Label-Stress Event Triggered Type (CL-H): users of this type have recently experienced a major stressful event, and the system already has a clear professional diagnostic label. AI text analysis indicates a high level of emotional activation, requiring rapid connection to manual consultation or crisis intervention channels. Silent Type (CL-I): users of this type rarely interact and lack effective behavioral or emotional data. Such users should be activated through gentle content push and passive guidance mechanisms. In actual application, for the current user, the cluster category of the current user can be determined according to the user characteristic vector, and then the corresponding preset second intervention strategy is assigned to the current user to perform psychological intervention and obtain an intervention effect label.

[0034] In one specific embodiment, classification model training can first determine user group patterns (i.e., cluster labels) through clustering. A prediction model is then trained based on historical intervention effects. Based on user behavioral characteristics and psychological states (both immediate and short-term effect tracking), the intervention effect label (excellent / good / average / poor / harmful) is determined. If the intervention effect label is "excellent," the current intervention plan is recommended with increased intensity. If the intervention effect label is "harmful," the current intervention plan is immediately stopped and an alternative plan is switched. Other intervention plans can be observed and adjusted. Training data includes user feature vectors, cluster labels, a pre-set second intervention strategy, and effect labels. The classification model uses user feature vectors, cluster labels (for users), and the pre-set second intervention strategy as inputs during prediction and iteration, and outputs the effect (effect label). The model can be periodically retrained using newly labeled data, and cluster boundaries and classification decision boundaries can be adjusted (increasing or decreasing cluster labels) and classification decision boundaries (increasing or decreasing effect labels). Incremental learning (new effect feedback data) can be used to update the classification model.

[0035] In a specific embodiment, the preset second intervention strategy may include scale tools, psychological self-help tools, psychology popular science courses, AI chat frequency, and therapy skills. For primary school users, the scale tools include: Children's Depression Inventory (CDI) pre-test and post-test; psychological self-help tools include: Little Rabbit's Rainy Day, Echo of the Sea, Adventures in the Emotional Planet, Beat the Thinking Monster, Emotional Adventure, I am a Treasure Child, and Look at Your Strengths; psychology popular science courses include: Kiki's Mood Weather Station, Positive Suggestion Change; AI chat frequency: mild 1 time / week, moderate 2-3 times / week, severe 4-5 times / week; therapy skills include: gamified CBT cognitive reconstruction, emotion recognition and expression. For junior high school users, the scale tools include: Patient Health Questionnaire (PHQ-9) pre-test and post-test; psychological self-help tools include: Echoes of the Sea, Adventures in the Planet of Emotions, Beat the Monsters of Thought, Adventures in Emotions, I am a Treasure Child, and Look at Your Strengths; psychology popular science courses include: Make Emotions More Sweet and Less Salty, Where is Your Emotional 'Control Panel', and Why Does Your Body Get Grumpy When You're in a Bad Mood; AI Chat Frequency: Mild 1 time / week, Moderate 2-3 times / week, Severe 4-5 times / week; Therapeutic Techniques: CBT Cognitive Restructuring, Behavioral Activation, and Mindfulness Meditation. For high school and adult users, the scale tools include: Patient Health Questionnaire (PHQ-9) pre-test and post-test; psychological self-help tools include: Echoes of the Sea, Adventures in the Emotional Planet, Beat the Thinking Monster, Emotional Adventures, I am a Treasure Child, and Look at Your Strengths; Psychology Popular Science Courses: What to Do If You Always Can't Get Along with Yourself, and Adventures in Youth Emotions; Frequency of AI Chat: Mild 1 time / week, Moderate 2-3 times / week, Severe 4-5 times / week; Therapy Techniques: Deep CBT, ACT Acceptance and Commitment, and DBT Emotional Regulation.

[0036] In some embodiments, using the same preset second intervention strategy to perform psychological intervention on users in the same cluster category to obtain intervention effect labels includes: Utilize the same preset second intervention strategy to conduct psychological intervention on users in the same cluster category, track the immediate effect and short-term effect of the intervention, obtain the immediate effect score and short-term effect score respectively, determine the comprehensive effect score based on the immediate effect score and the short-term effect score, and obtain the intervention effect label based on the comprehensive effect score.

[0037] In a specific embodiment, the comprehensive effect score = α × immediate effect score + β × short-term effect score. The intervention effect label can be divided into five levels according to the comprehensive effect score. Excellent is the first level (comprehensive effect score ≥ 1.5 points), indicating that the intervention effect is significant and the user benefits greatly. Good is the second level (comprehensive effect score is [1.0, 1.5) points), indicating that the intervention effect is obvious and the user benefits significantly. Medium is the third level (comprehensive effect score is [0.5, 1.0) points), indicating that the intervention effect is general and the user benefits partially. Poor is the fourth level (comprehensive effect score is [0-0.5) points), indicating that the intervention effect is weak and the user benefits slightly. Harmful is the fifth level (comprehensive effect score is <0 points), indicating that the intervention has a negative effect. If the level of the intervention effect label is lower than the preset label level (for example, the third level), the preset second intervention strategy is stopped from being pushed to the user.

[0038] In some embodiments, the immediate effect tracking includes monitoring mood changes, platform engagement, behavioral responses, and subjective satisfaction, and the short-term effect tracking includes monitoring symptom improvement effects, functional recovery effects, treatment compliance, and side effects.

[0039] In a specific embodiment, immediate effect tracking includes monitoring emotional changes, platform engagement, behavioral responses and subjective satisfaction, which can be quantified with a numerical value of [0,1]. Immediate effect tracking can be effect tracking for 0-24 hours. Short-term effect tracking includes monitoring symptom improvement effects, functional recovery effects, treatment compliance and side effects, which can also be quantified with a numerical value of [0,1]. Short-term effect tracking can be effect tracking for 1-4 weeks.

[0040] The mental health intervention method provided by the embodiment of the present invention obtains the user's evaluation behavior characteristic data, the evaluation behavior characteristic data includes the answer consistency score, the completion time rationality score and the response pattern regularity score; determines the credibility index according to the evaluation behavior characteristic data, determines the warning level according to the credibility index, and determines the preset first intervention strategy according to the warning level; after the preset first intervention strategy acts on the user, obtains the user characteristic data, and forms a user characteristic vector according to the user characteristic data; determines the user's clustering category according to the user characteristic vector, and assigns the corresponding preset second intervention strategy to the user through the trained classification model and the user's clustering category to perform psychological intervention and obtain an intervention effect label. If the level of the intervention effect label is lower than the preset label level, the preset second intervention strategy is stopped from being pushed to the user; the classification model is trained based on the user characteristic vector sample and the preset second intervention strategy; it can realize early warning and intervention of mental health, improve the rationality of mental health intervention, and thus improve the effect of mental health intervention.

[0041] The mental health intervention method provided by the embodiment of the present invention can achieve closed-loop management of the entire process from initial assessment and evaluation to continuous intervention. By deeply mining the user's all-round behavioral data, dynamically optimizing the intervention strategy, and integrating the user's assessment data, behavioral data, feedback data and other multi-dimensional information, it can achieve closed-loop optimization of dynamic risk warning, intelligent intervention recommendation and effect tracking feedback, realize truly personalized and adaptive continuous and precise intervention, and improve the effect of mental health intervention.

[0042] Based on the above-mentioned mental health intervention method, an embodiment of the present invention provides a mental health intervention device, the structural diagram of which is as follows: Figure 2 As shown, the mental health intervention device 20 includes a data acquisition module 21, a warning level determination module 22, a characteristic vector generation module 23 and an intervention module 24; The data acquisition module 21 is used to obtain the user's evaluation behavior characteristic data, which includes the answer consistency score, the completion time rationality score and the response pattern regularity score; The warning level determination module 22 is configured to determine a credibility index based on the assessed behavior characteristic data, determine a warning level based on the credibility index, and determine a preset first intervention strategy based on the warning level; The characteristic vector generating module 23 is configured to obtain user characteristic data after the preset first intervention strategy acts on the user, and form a user characteristic vector based on the user characteristic data; The intervention module 24 is used to determine the user's cluster category based on the user characteristic vector, and assign the corresponding preset second intervention strategy to the user through the trained classification model and the user's cluster category to perform psychological intervention and obtain an intervention effect label. If the level of the intervention effect label is lower than the preset label level, the preset second intervention strategy is stopped from being pushed to the user; the classification model is trained based on the user characteristic vector sample and the preset second intervention strategy.

[0043] For other details about how the modules in the mental health intervention device implement the above technical solutions, please refer to the description of the mental health intervention method provided in the above invention embodiments, which will not be repeated here.

[0044] Figure 3 Schematic diagram of the structure of an electronic device according to an embodiment of the present invention. Figure 3 As shown, the electronic device 30 includes a processor 31 and a memory 32 communicatively connected to the processor 31 .

[0045] The memory 32 stores program instructions for implementing the mental health intervention method of any one of the above embodiments.

[0046] The processor 31 is configured to execute program instructions stored in the memory 32 to perform mental health intervention.

[0047] The processor 31 may also be referred to as a CPU (Central Processing Unit). The processor 31 may be an integrated circuit chip having signal processing capabilities. The processor 31 may also be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic device, a discrete gate or transistor logic device, or a discrete hardware component. The general-purpose processor may be a microprocessor or any conventional processor.

[0048] An embodiment of the present invention provides a computer-readable storage medium, and the storage medium of the embodiment of the present invention stores program instructions that can implement all the above methods, and the storage medium can be non-volatile or volatile. Among them, the program instructions can be stored in the above storage medium in the form of a software product, including a number of instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) or a processor to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, or terminal devices such as a computer, a server, a mobile phone, and a tablet.

[0049] In the several embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of modules is merely a logical function division. In actual implementation, there may be other division methods, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be through some interface, indirect coupling or communication connection of devices or modules, which can be electrical, mechanical or other forms.

[0050] In addition, the functional modules in the various embodiments of the present invention may be integrated into one processing unit, or each module may exist physically separately, or two or more modules may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of a software functional unit. The above is only an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

[0051] The above description is only an embodiment of the present invention. It should be pointed out that those skilled in the art can make improvements without departing from the creative concept of the present invention, but these improvements all fall within the scope of protection of the present invention.

Claims

1. A mental health intervention method, characterized in that: include: Obtaining user assessment behavior characteristic data, wherein the assessment behavior characteristic data includes a consistency score for answering questions, a reasonableness score for completion time, and a regularity score for response patterns; Determining a credibility index based on the assessed behavioral characteristic data, determining a warning level based on the credibility index, and determining a preset first intervention strategy based on the warning level; After the preset first intervention strategy acts on the user, obtaining user feature data, and forming a user characteristic vector according to the user feature data; The user's clustering category is determined based on the user's characteristic vector. Through the trained classification model and the user's clustering category, the corresponding preset second intervention strategy is assigned to the user to perform psychological intervention and obtain an intervention effect label. If the level of the intervention effect label is lower than the preset label level, the preset second intervention strategy is stopped from being pushed to the user; the classification model is trained based on the user's characteristic vector sample and the preset second intervention strategy.

2. The mental health intervention method according to claim 1, characterized in that: Determining a credibility index based on the assessed behavior characteristic data, and determining a warning level based on the credibility index, including: A weighted sum is performed on the answer consistency score, the completion time rationality score and the response pattern regularity score to obtain a credibility index. If the credibility index is not greater than the first credibility threshold, the warning level is determined to be the first level; if the credibility index is greater than the first credibility threshold and not greater than the second credibility threshold, the warning level is determined to be the second level; if the credibility index is greater than the second credibility threshold and not greater than the third credibility threshold, the warning level is determined to be the third level; if the credibility index is greater than the third credibility threshold and not greater than the fourth credibility threshold, the warning level is determined to be the fourth level; if the credibility index is greater than the fourth credibility threshold, the warning level is determined to be the fifth level.

3. The mental health intervention method according to claim 2, characterized in that: Determining a preset first intervention strategy according to the warning level includes: If the warning level is level one, the preset first intervention strategy is regular monitoring; if the warning level is level two, the preset first intervention strategy is the use of a self-help resource package; if the warning level is level three, the preset first intervention strategy is the use of a combination of artificial intelligence consultation and self-help tools; if the warning level is level four, the preset first intervention strategy is the use of professional psychological counselors and auxiliary tools; if the warning level is level five, the preset first intervention strategy is emergency medical referral and immediate psychological intervention.

4. The mental health intervention method according to claim 1, characterized in that: Acquiring user feature data and forming a user characteristic vector based on the user feature data includes: Obtain course learning and interactive behavior dimension data, self-service tool usage and interaction dimension data, artificial intelligence consultation multimodal interaction dimension data, and professional consultation feedback integration dimension data; use at least one dimension data among the course learning and interactive behavior dimension data, the self-service tool usage and interaction dimension data, the artificial intelligence consultation multimodal interaction dimension data, and the professional consultation feedback integration dimension data as user feature data, and form a user feature vector based on the user feature data.

5. The mental health intervention method according to claim 1, characterized in that: The steps of training the classification model include: Clustering the user feature vector samples to determine the cluster categories of the corresponding users, using the same preset second intervention strategy to perform psychological intervention on users in the same cluster category, obtaining intervention effect labels, and optimizing the intervention effect labels by adjusting the preset second intervention strategy.

6. The mental health intervention method according to claim 5, characterized in that: Performing psychological intervention on users in the same cluster category using the same preset second intervention strategy to obtain intervention effect labels, including: Utilize the same preset second intervention strategy to conduct psychological intervention on users in the same cluster category, track the immediate effect and short-term effect of the intervention, obtain the immediate effect score and short-term effect score respectively, determine the comprehensive effect score based on the immediate effect score and the short-term effect score, and obtain the intervention effect label based on the comprehensive effect score.

7. The mental health intervention method according to claim 6, characterized in that: The immediate effect tracking includes monitoring mood changes, platform engagement, behavioral responses and subjective satisfaction, and the short-term effect tracking includes monitoring symptom improvement effects, functional recovery effects, treatment compliance and side effects.

8. A mental health intervention device, characterized in that: It includes data acquisition module, warning level determination module, characteristic vector generation module and intervention module; The data acquisition module is used to obtain the user's evaluation behavior characteristic data, wherein the evaluation behavior characteristic data includes the answer consistency score, the completion time rationality score and the response pattern regularity score; The warning level determination module is configured to determine a credibility index based on the assessed behavior characteristic data, determine a warning level based on the credibility index, and determine a preset first intervention strategy based on the warning level; The characteristic vector generating module is configured to obtain user characteristic data after the preset first intervention strategy acts on the user, and form a user characteristic vector based on the user characteristic data; The intervention module is used to determine the user's cluster category based on the user characteristic vector, and assign the corresponding preset second intervention strategy to the user through the trained classification model and the user's cluster category to perform psychological intervention and obtain an intervention effect label. If the level of the intervention effect label is lower than the preset label level, the preset second intervention strategy is stopped from being pushed to the user; the classification model is trained based on the user characteristic vector sample and the preset second intervention strategy.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program executable by the processor, wherein: When the processor executes the computer program, the mental health intervention method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the mental health intervention method according to any one of claims 1 to 7 is implemented.

Images