College student psychological health service system and method based on AI Agent

Abstract

The invention relates to the field of intelligent psychological health management service, in particular to an AI Agent-based college student psychological health service system and method, and the system comprises a psychological health measurement module, a social personality portrait module and a psychological health intervention module. According to the method, a measurer agent measures the current psychological health condition of a college student from a transverse angle through a standardized psychological health scale; the portrayman agent analyzes the historical psychological health conditions and MBTI personality portraits of the college student users from the longitudinal angle by using the social media data of the college student users obtained by the web crawler; based on psychological health scale data and social media data, the system adaptively recommends a consultant agent for college students, the consultant agent is externally connected with a psychological consultation expert guide and real consultation case data, professional and effective psychological health intervention services are provided for the college students, and the system is suitable for college student emotion accompanying, college auxiliary decision making and other scenes.

Description

Technical Field

[0001] The present invention relates to the field of intelligent mental health management services, and in particular to a system and method for providing mental health services to college students based on an AI agent with a large model. Background Art

[0002] The core contents of mental health management include mental health information collection, mental health risk assessment and mental health intervention. It often uses mental health scales to measure the mental health status of individuals, uses social media data to create social personality portraits, and conducts targeted mental health interventions based on mental health scale data and social media data.

[0003] Among them, the AI ​​Agent-based college student mental health service system refers to a mental health measurement and intervention system built with the help of large-scale model-based intelligent agent technology. The specific methods include the surveyor intelligent agent using a standardized mental health scale to measure the current mental health status of college students from a horizontal perspective; the portraitur intelligent agent uses the social media data of college students obtained by web crawlers to analyze their historical mental health status and MBTI personality portrait from a vertical perspective; the system adaptively recommends counselor intelligent agents to college students based on mental health scale data and social media data. The counselor intelligent agent is connected to the psychological counseling expert guide and real counseling case data to provide professional and effective mental health intervention services to college students.

[0004] In the process of realizing the present invention, the inventors found that there are at least the following problems in the existing technology: First, there is a lack of professionalism. Most mental health service systems lack the support of professional knowledge in psychology, and the services are superficial emotional support; second, there is a lack of targeting. The system ignores the heterogeneity of college students' mental health problems and adopts a "one-size-fits-all" standardized response; third, it lacks attractiveness. The mental health reports provided by the system are poorly readable, and the interactive form is limited to text. Summary of the Invention

[0005] The purpose of this invention is to solve the shortcomings of the existing technology and propose an AI Agent-based college student mental health service system and method.

[0006] To achieve the above objectives, the present invention adopts the following technical solutions: an AI Agent-based college student mental health service system and method comprising:

[0007] In the mental health measurement module, the surveyor agent calculates the multidimensional mental health scores of college students using a standardized mental health scale and generates multiple mental health analysis reports, a mental health radar chart, a mental health comprehensive report, and a mental health animal portrait.

[0008] In the social personality profiling module, the profiling agent uses the social media data of college students obtained by web crawlers to generate multiple descriptive statistics reports, a sentiment analysis report, a Myers-Briggs Type Indicator (MBTI) personality prediction report, and an MBTI personality profile.

[0009] In the mental health intervention module, the counselor agent is connected to the psychological counseling expert guide and real counseling case data. Based on the mental health measurement module data and the social personality portrait module data, the system adaptively recommends counselor agents for college students. The counselor agent provides professional and effective mental health intervention services to college students.

[0010] As a further solution of the present invention, the standardized mental health scale includes a general health scale, a happiness index scale and a perceived stress scale; the descriptive statistical report includes a post quantity report, a post time report, a post word cloud report and a seasonal word cloud report; the psychological counseling expert guide includes cognitive behavioral therapy, solution-focused brief therapy, positive psychotherapy and emotion-focused therapy.

[0011] As a further solution of the present invention, the mental health measurement module includes:

[0012] The mental health score calculation submodule obtains the mental health scale data of college students. The general health scale score calculation formula is: G=(9-S)+D+A, where Q i ∈{0, 1}, corresponding to "no" and "yes" respectively, S is the score of the self-affirmation dimension, D is the score of the depression dimension, A is the score of the anxiety dimension, and G is the total score of the general health scale; the calculation formula of the happiness index scale score is: L = Q9, H = 0 / 8 × 1 + L × 1.1, where Q j , Q9∈{1, 2, ..., 7}, O is the overall emotional index dimension score, L is the life satisfaction index dimension score, and H is the total score of the happiness index scale; the calculation formula for the perceived stress scale score is: T=∑ k∈{1，2，3，8，11，12，14} Q k , C=∑ k∈{4,5，6，7，9，10，13} (4-Q k ), P=T+C, where Q k ∈{0, 1, 2, 3, 4}, T is the score of tension dimension, C is the score of loss of control dimension, and P is the total score of perceived stress scale;

[0013] The mental health analysis report submodule uses large-scale model-based prompt word engineering technology, based on the overview and scoring criteria of the three mental health scales and the structure of conventional mental health analysis reports. It combines the three mental health scale data of college students to generate composite prompt words. The surveyor agent then outputs multiple mental health analysis reports and a mental health radar chart.

[0014] The mental health comprehensive report submodule uses large-scale model-based prompt word engineering technology, based on the conventional mental health comprehensive report structure, combined with multiple mental health analysis reports of college students, to generate composite prompt words, and then the surveyor agent outputs a mental health comprehensive report;

[0015] The mental health portrait sub-module uses the large model's prompt word engineering technology and multimodal output technology, based on the overview of Plutchik's emotion wheel and the content of conventional mental health animal portraits, combined with the comprehensive mental health report of college students, to combine and generate compound prompt words, and then the surveyor intelligent agent outputs a mental health animal portrait.

[0016] As a further solution of the present invention, the social personality portrait module includes:

[0017] The web crawler module obtains the uniform resource locators of the posts through the social media accounts provided by college students, and then performs multimodal web crawling to obtain the social media data of college students;

[0018] The descriptive statistics submodule uses large-scale model-based prompt word engineering technology and word cloud graph technology, based on the structure of conventional social media descriptive statistics reports and combined with the social media data of college students to generate compound prompt words. The portrait artist agent then outputs multiple descriptive statistics reports.

[0019] After systematically processing the social media data of college students, the sentiment analysis submodule uses large-scale model-based prompt word engineering technology and SnowNLP sentiment analysis technology. Based on the structure of conventional social media sentiment analysis reports, it combines the social media data of college students to generate compound prompt words. The portrait artist agent then outputs a sentiment analysis report.

[0020] The MBTI personality analysis submodule leverages the large model's prompt word engineering technology and multimodal output technology, based on the MBTI scale's overview, scoring criteria, and conventional personality analysis report structure, combined with the social media data of college students, to generate composite prompt words. The portrait artist agent then outputs an MBTI personality inference report and an MBTI personality portrait.

[0021] As a further embodiment of the present invention, the mental health intervention module includes:

[0022] The database submodule connects the counselor's intelligent body to the psychological counseling expert guide and real counseling case data. It uses prompt word engineering technology to connect the counselor's intelligent body to the psychological counseling expert guide. It also extracts information through prompt word engineering technology to obtain a keyword list. This keyword list is then vectorized using a vectorized embedding model and stored in a vector library, connecting the counselor's intelligent body to real counseling case data.

[0023] The adaptive recommendation submodule leverages the large-scale model's prompt word engineering technology, based on the overview of the psychological counseling multidimensional matching model and the structure of conventional counselor recommendation reports. It combines data from the college students' mental health measurement module and social personality profiling module to generate composite prompt words, and then outputs a counselor agent recommendation report.

[0024] The psychological counseling service sub-module uses the large model's prompt word engineering technology, retrieval enhancement generation technology and thinking chain technology, based on the psychological counseling expert guide and real counseling case data, combined with the college students' mental health measurement module data and social personality portrait module data, to combine and generate compound prompt words, and then the counselor intelligent body provides professional and effective mental health intervention services to college students.

[0025] The AI ​​Agent-based mental health service method for college students includes the following steps:

[0026] S1: The surveyor agent calculates the multidimensional mental health scores of college students using a standardized mental health scale;

[0027] S2: The surveyor agent leverages the large-scale model's prompt word engineering and multimodal output technology to generate multiple mental health analysis reports, a mental health radar chart, a mental health summary report, and a mental health animal portrait.

[0028] S3: The portrait agent uses the social media data of college students obtained by web crawlers to generate multiple descriptive statistics reports and a sentiment analysis report;

[0029] S4: The portrait artist agent uses the large model's prompt word engineering technology and multimodal output technology to generate an MBTI personality prediction report and an MBTI personality portrait;

[0030] S5: The counselor agent is connected to the psychological counseling expert guide and real counseling case data. Based on the data of the mental health measurement module and the social personality profile module, the system adaptively recommends counselor agents for college students.

[0031] S6: The counselor intelligent agent uses the large model's prompt word engineering technology, retrieval enhancement generation technology and thinking chain technology to provide professional and effective mental health intervention services to college students.

[0032] Compared with the prior art, the advantages and positive effects of the present invention are:

[0033] In the present invention, the system is built based on psychological knowledge such as the Mental Health Questionnaire, Plutchik's Emotional Wheel, MBTI Scale, Psychological Counseling Expert Guide, real counseling case data and psychological counseling multidimensional matching model to ensure sufficient professionalism; the system uses prompt word engineering, word cloud diagrams, sentiment analysis, retrieval enhancement generation and thinking chain technologies to improve the pertinence of services; the system uses multimodal output, user-friendly front-end and other designs to provide vivid and interesting analysis reports. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 is a system flow chart of the present invention;

[0035] Figure 2 This is a flow chart of the mental health measurement module of the present invention;

[0036] Figure 3 This is a flow chart of the social personality portrait module of the present invention;

[0037] Figure 4 This is a flow chart of the mental health intervention module of the present invention. DETAILED DESCRIPTION

[0038] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0039] In the description of the present invention, it should be understood that the term "plurality" means two or more than two, unless otherwise clearly defined.

[0040] See also Figure 1 , the AI ​​Agent-based college student mental health service system includes:

[0041] In the mental health measurement module, the surveyor agent calculates the multidimensional mental health scores of college students using a standardized mental health scale and generates multiple mental health analysis reports, a mental health radar chart, a mental health comprehensive report, and a mental health animal portrait.

[0042] In the social personality profiling module, the profiling agent uses the social media data of college students obtained by web crawlers to generate multiple descriptive statistical reports, a sentiment analysis report, an MBTI personality prediction report, and an MBTI personality profile.

[0043] In the mental health intervention module, the counselor agent is connected to the psychological counseling expert guide and real counseling case data. Based on the mental health measurement module data and the social personality portrait module data, the system adaptively recommends counselor agents for college students. The counselor agent provides professional and effective mental health intervention services to college students.

[0044] See also Figure 2 , the mental health measurement module includes:

[0045] The mental health score calculation submodule obtains the mental health scale data of college students. The general health scale score calculation formula is: G=(9-S)+D+A, where Q i ∈{0, 1}, corresponding to "no" and "yes" respectively, S is the score of the self-affirmation dimension, D is the score of the depression dimension, A is the score of the anxiety dimension, and G is the total score of the general health scale; the calculation formula of the happiness index scale score is: L=Q9,H=O / 8×1+L×1.1,where Q j , Q9∈{1, 2, ..., 7}, O is the overall emotional index dimension score, L is the life satisfaction index dimension score, and H is the total score of the happiness index scale; the calculation formula for the perceived stress scale score is: T=∑ k∈{1，2，3，8，11，12，14} Q k , C=∑ k∈{4,5，6，7，9，10,13} (4-Q k ), P=T+C, where Q k ∈{0, 1, 2, 3, 4}, T is the score of the tension dimension, C is the score of the sense of loss of control dimension, and P is the total score of the perceived stress scale; assuming that a college student answers "no" when filling out the general health scale, fills 2 points for questions 1-4 and 5 points for questions 5-9 when filling out the happiness index scale, and fills 4 points when filling out the perceived stress scale, then the data of the three mental health scales are as follows: the general health scale score is S=1+1+1+1+1+1+0+1+1=8, D=0+1+1+1+1+1=5 , A=1+1+1+1+1=5, G=(9-8)+5+5=11; the happiness index scale score is O=2+2+2+2+4+4+4+4=24, L=4, H=24 / 8×1+4×1.1=7.4; the perceived stress scale score is T=4+4+4+4+4+4+4=28, C=(4-4)+(4-4)+(4-4)+(4-4)+(4-4)+(4-4)+(4-4)=0, P=28+0=28.

[0046] The mental health analysis report submodule uses large-scale model-based prompt word engineering technology, based on the overview and scoring criteria of the three mental health scales and the structure of conventional mental health analysis reports. It combines the three mental health scale data of college students to generate composite prompt words. The surveyor agent then outputs multiple mental health analysis reports and a mental health radar chart.

[0047] In the general health scale scores, higher scores on the self-affirmation dimension indicate higher levels of self-affirmation, higher scores on the depression dimension indicate higher levels of depression, higher scores on the anxiety dimension indicate higher levels of anxiety, and higher total scores on the general health scale indicate worse mental health conditions of the individual; in the happiness index scale scores, higher scores on the overall emotion index dimension indicate more positive overall emotions, higher scores on the life satisfaction index dimension indicate higher life satisfaction, and higher total scores on the happiness index scale indicate higher individual happiness; in the stress perception scale scores, higher scores on the tension dimension indicate stronger tension, higher scores on the sense of loss of control dimension indicate stronger sense of loss of control, and higher total scores on the perceived stress scale indicate greater perceived stress of the individual. The mental health radar chart includes five dimensions: self-denial, depression, anxiety, unhappiness index, and perceived stress, which correspond to the self-affirmation dimension, depression dimension, anxiety dimension, happiness index scale and perceived stress scale of the general health scale respectively; assuming that a college student S=8, D=5, A=5, G=11, O=24, L=4, H=7.4, T=28, C=0, P=28, input the overview of the general health scale, scoring criteria, conventional mental health analysis report structure and the general health scale score of the college student S=8, D=5, A=3, G=11, and the surveyor intelligent agent outputs a mental health analysis report including the overall general health situation, self-affirmation dimension, depression dimension and anxiety dimension; input the overview of the happiness index scale, The scoring criteria, the structure of the conventional mental health analysis report and the happiness index score of college students are O=24, L=4, H=7.4, and the surveyor intelligent agent outputs a mental health analysis report including the overall happiness index, the overall emotional index dimension and the life satisfaction index dimension; the overview of the perceived stress scale, the scoring criteria, the structure of the conventional mental health analysis report and the perceived stress scale score of college students are T=28, C=0, P=28, and the surveyor intelligent agent outputs a mental health analysis report including the overall perceived stress, the tension dimension and the sense of loss of control dimension; the scale score S=8, D=5, A=5, H=7.4, P=28 filled in by college students is input, and the surveyor intelligent agent outputs a five-dimensional mental health radar chart.

[0048] The mental health comprehensive report sub-module uses the large model-based prompt word engineering technology, based on the conventional mental health comprehensive report structure, combined with multiple mental health analysis reports of college students, to generate compound prompt words, and then the surveyor intelligent agent outputs a mental health comprehensive report; assuming that multiple mental health analysis reports of a college student have been generated, the conventional mental health comprehensive report structure and multiple mental health analysis reports of college students are input, and the surveyor intelligent agent outputs a mental health comprehensive report including the overall mental health situation and mental health improvement suggestions.

[0049] The mental health portrait submodule leverages the large-scale model's prompt word engineering and multimodal output technology, based on the overview of Plutchik's emotion wheel and the content of conventional mental health animal portraits, combined with the university students' comprehensive mental health reports, to generate composite prompt words. The surveyor agent then outputs a mental health animal portrait.

[0050] The Plutchik Emotion Wheel corresponds to yellow for happiness, light green for trust, dark green for fear, light blue for surprise, dark blue for sadness, purple for disgust, red for anger, and orange for expectation. Suppose a comprehensive mental health report for a college student has been generated, and the overview of the Plutchik Emotion Wheel, the content of conventional mental health animal portraits, and the comprehensive mental health report of the college student are input. If the overall mental health of the college student is happy, the surveyor agent may output a yellow lion.

[0051] See also Figure 3 , the social personality portrait module includes:

[0052] The web crawler submodule obtains the uniform resource locator of the posts through the social media accounts provided by college students, and then performs multimodal web crawlers to obtain the social media data of college students; assuming that a college student chooses the Weibo platform and his account name is "xx", the system will automatically go to Weibo to search for his account, enter the college student's personal homepage, obtain the uniform resource locator of the posts on the personal homepage, and then the system crawls the text, pictures, videos and other content in the posts, and stores them in the college student's social media database after data cleaning.

[0053] The descriptive statistics submodule uses the prompt word engineering technology and word cloud diagram technology based on the large model, takes the conventional social media descriptive statistics report structure as the basis, combines the social media data of college students, and generates compound prompt words, and then the portrait artist agent outputs multiple descriptive statistics reports; assuming that a social media database of a college student user has been established, the conventional post quantity report structure and the social media data of the college student user are input, and the portrait artist agent outputs a descriptive statistics report including the number of posts, forwardings, comments, likes, texts, pictures and videos of the college student user; the conventional posting time ... Output a descriptive statistical report including the changes in the number of posts, the number of reposts, the number of comments and the number of likes over time, as well as the posting situation in the early morning; input the regular posting word cloud report structure and the social media data of college students, and the portrait artist intelligent agent outputs a descriptive statistical report that extracts the keywords of college students' posts; posts from January to March are spring posts, posts from April to June are summer posts, posts from July to September are autumn posts, and posts from October to December are winter posts. Input the regular season word cloud report structure and the social media data of college students, and the portrait artist intelligent agent outputs a descriptive statistical report including the word cloud map of spring posts, the word cloud map of summer posts, the word cloud map of autumn posts and the word cloud map of winter posts.

[0054] After systematically processing the social media data of college students, the sentiment analysis submodule uses large-scale model-based prompt word engineering technology and SnowNLP sentiment analysis technology. Based on the structure of conventional social media sentiment analysis reports, it combines the social media data of college students to generate compound prompt words. The portrait artist agent then outputs a sentiment analysis report.

[0055] Systematic processing includes noise cleaning, post text standardization, word segmentation and stop word filtering. SnowNLP sentiment analysis technology is often used for sentiment tendency calculation. By comprehensively considering the sentiment polarity, negation structure, degree adverbs and syntactic features of college students' social media data, the sentiment of their posts is quantitatively evaluated to obtain a post sentiment score S in the interval [0, 1], where S≈1 represents highly positive emotions, S≈0 represents highly negative emotions, and S≈0.5 represents neutral or ambiguous emotions. Suppose a college student user posts a total of 5 posts. After systematic processing, the sentiment tendency is calculated using SnowNLP sentiment analysis technology, and the post sentiment score S is {0.37, 0.12, 0.35, 0.56, 0.78}. The portrait artist intelligent agent then outputs a sentiment analysis report including the most positive post S=0.78 and the most negative post S=0.12.

[0056] The MBTI personality analysis submodule leverages the large-scale model's prompt word engineering and multimodal output technology. Based on the MBTI scale's overview, scoring criteria, and conventional personality analysis report structure, combined with university students' social media data, it generates composite prompt words. The profiler agent then outputs an MBTI personality inference report and an MBTI personality portrait.

[0057] MBTI personality is divided into sixteen types based on the four dimensions of Extraversion (E) and Introversion (I), Sensing (S) and Intuition (N), Thinking (T) and Feeling (F), and Judging (J) and Perceiving (P). The MBTI scale overview, scoring criteria, conventional personality inference report structure, and the social media data of college students are input, and the portrait artist agent outputs an MBTI personality inference report. MBTI personality portraits are divided into four colors based on the two dimensions of S and N, and T and F, and each color includes four personality types. The MBTI scale overview, scoring criteria, conventional personality portrait content, and the social media data of college students are input, and the portrait artist agent outputs an MBTI personality portrait with gender. Suppose a college student user has published a total of 23 posts. After all posts are input into the large model, the portrait artist agent outputs an MBTI personality inference report inferring ENFP, with N and F corresponding to green. Since the college student user's social media data indicates that she is female, the portrait artist agent outputs a green female MBTI personality portrait.

[0058] See also Figure 4 , mental health intervention modules include:

[0059] The database submodule connects the counselor's intelligent body to the psychological counseling expert guide and real counseling case data. It uses prompt word engineering technology to connect the counselor's intelligent body to the psychological counseling expert guide. It also extracts information through prompt word engineering technology to obtain a keyword list. This keyword list is then vectorized using a vectorized embedding model and stored in a vector library, connecting the counselor's intelligent body to real counseling case data.

[0060] The expert guide to psychological counseling is divided into four schools of psychological counseling: cognitive behavioral therapy, solution-focused brief therapy, positive psychotherapy, and emotion-focused therapy. Cognitive behavioral therapy includes four steps: identifying automatic thoughts, identifying thought traps, challenging automatic thoughts, and returning to realistic thinking. Solution-focused brief therapy includes five steps: problem description, goal setting, exploring exceptions, counseling feedback, and evaluating progress. Positive psychotherapy includes fifteen steps: positive introduction and gratitude journal, character strengths and significant strengths, practical wisdom, a better self, open memory and closed memory, forgiveness, maximization and satisfaction, gratitude, hope and optimism, post-traumatic growth, slow living and enjoying life, positive relationships, positive communication, altruistic behavior, and the meaning and purpose of life. Emotion-focused therapy includes four steps: building relationships and initial assessment, exploring and reaching core emotions, emotional transformation and reconstruction, and consolidation and closing counseling.

[0061] The real consultation case data includes more than 200 psychological counseling streaming videos, covering multiple topics and various psychological counseling schools. The real consultation case data is divided into document blocks, and the types of document blocks are divided into three categories: text, pictures, and tables. Among them, pictures are written in a format based on 64 printable characters to represent binary data, and tables are written in plain text format. At the same time, extra-long tables are divided and stored twice. If there are captions above and below the picture or table, the captions are merged with the picture or table information.

[0062] The adaptive recommendation submodule uses the prompt word engineering technology of the large model, based on the overview of the multidimensional matching model of psychological counseling and the structure of the recommendation report of conventional counselors, combined with the mental health measurement module data and social personality portrait module data of college students, to generate compound prompt words, and then output a counselor intelligent agent recommendation report; assuming that the mental health measurement module and social personality portrait module find that a college student has more serious anxiety, the overview of the multidimensional matching model of psychological counseling, the structure of the recommendation report of conventional counselors and the mental health measurement module data and social personality portrait module data of college students are input, then a counselor intelligent agent recommendation report recommending emotion-focused therapy may be output.

[0063] The psychological counseling service submodule uses the large model's prompt word engineering technology, retrieval enhancement generation technology and thinking chain technology, based on the psychological counseling expert guide and real counseling case data, combined with the college students' mental health measurement module data and social personality portrait module data, to generate compound prompt words, and then the counselor agent provides professional and effective mental health intervention services to college students; suppose a college student inputs to the emotion-focused therapy counselor agent: "There are so many homeworks and so much pressure", first, the counselor agent extracts the keywords input by the college student: "homework, pressure, big"; secondly, the counselor agent traverses the emotion-focused therapy expert guide based on the extracted keywords, and selects the keywords from the expert guide The appropriate link is selected: "Link 1: Establishing a relationship and preliminary assessment", and through the thinking chain technology, reasoning is carried out to preliminarily conceive the counseling suggestions to be provided to college students: "College students express that they are under great pressure and have a lot of homework. First, establish a trusting relationship, listen carefully and understand the feelings of college students, and through questions and feedback, help college students clarify that the main problem they are facing is excessive academic pressure. Preliminary assessment of college students' emotional state and coping methods, understand college students' experience of emotions, and pave the way for subsequent in-depth exploration and intervention"; thirdly, the counselor intelligent agent uses retrieval enhancement generation technology to retrieve extracted keywords in real counseling case data and obtain case slices: "EMILY, You know, didn't piss anyone off...", and polishes the preliminarily conceived counseling suggestions based on the slices; finally, the counselor intelligent agent uses the psychological health measurement module data and social personality portrait module data of college students to optimize the current counseling suggestions in a targeted manner: "I can understand how you feel now. Facing heavy academic workload does make people feel very anxious and stressed. We can talk together and see how to make things easier."

[0064] The AI ​​Agent-based mental health service method for college students includes the following steps:

[0065] S1: The surveyor agent calculates the multidimensional mental health scores of college students using a standardized mental health scale;

[0066] S2: The surveyor agent leverages the large-scale model's prompt word engineering and multimodal output technology to generate multiple mental health analysis reports, a mental health radar chart, a mental health summary report, and a mental health animal portrait.

[0067] S3: The portrait agent uses the social media data of college students obtained by web crawlers to generate multiple descriptive statistics reports and a sentiment analysis report;

[0068] S4: The portrait artist agent uses the large model's prompt word engineering technology and multimodal output technology to generate an MBTI personality prediction report and an MBTI personality portrait;

[0069] S5: The counselor agent is connected to the psychological counseling expert guide and real counseling case data. Based on the data of the mental health measurement module and the social personality profile module, the system adaptively recommends counselor agents for college students.

[0070] S6: The counselor intelligent agent uses the large model's prompt word engineering technology, retrieval enhancement generation technology and thinking chain technology to provide professional and effective mental health intervention services to college students.

[0071] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. The AIAgent-based college student mental health service system is characterized by: The system comprises: In the mental health measurement module, the surveyor agent calculates the multidimensional mental health scores of college students using a standardized mental health scale and generates multiple mental health analysis reports, a mental health radar chart, a mental health comprehensive report, and a mental health animal portrait. In the social personality profiling module, the profiling agent uses the social media data of college students obtained by web crawlers to generate multiple descriptive statistical reports, a sentiment analysis report, an MBTI personality prediction report, and an MBTI personality profile. In the mental health intervention module, the counselor agent is connected to the psychological counseling expert guide and real counseling case data. Based on the mental health measurement module data and the social personality portrait module data, the system adaptively recommends counselor agents for college students. The counselor agent provides professional and effective mental health intervention services to college students.

2. The AI ​​Agent-based college student mental health service system according to claim 1 is characterized in that: The standardized mental health scale includes a general health scale, a happiness index scale and a perceived stress scale; the descriptive statistical report includes a post quantity report, a post time report, a post word cloud report and a seasonal word cloud report; the psychological counseling expert guide includes cognitive behavioral therapy, solution-focused brief therapy, positive psychotherapy and emotion-focused therapy.

3. The AI ​​Agent-based college student mental health service system according to claim 1 is characterized in that: The mental health measurement module includes: The mental health score calculation submodule obtains the mental health scale data of college students. The general health scale score calculation formula is: G=(9-S)+D+A, where Q i ∈{0, 1}, corresponding to "no" and "yes" respectively, S is the score of the self-affirmation dimension, D is the score of the depression dimension, A is the score of the anxiety dimension, and G is the total score of the general health scale; the calculation formula of the happiness index scale score is: L = Q9, H = 0 / 8 × 1 + L × 1.1, where Q j , Q9∈{1, 2, ..., 7}, 0 is the overall emotion index dimension score, L is the life satisfaction index dimension score, and H is the total score of the happiness index scale; the calculation formula for the perceived stress scale score is: T=∑ k∈{1，2，3，8，11，12，14} Q k , C=∑ k∈{4，5，6，7，9，10，13} (4-Q k ), P=T+C, where Q k ∈{0, 1, 2, 3, 4}, T is the score of tension dimension, C is the score of loss of control dimension, and P is the total score of perceived stress scale; The mental health analysis report submodule uses large-scale model-based prompt word engineering technology, based on the overview and scoring criteria of the three mental health scales and the structure of conventional mental health analysis reports. It combines the three mental health scale data of college students to generate composite prompt words. The surveyor agent then outputs multiple mental health analysis reports and a mental health radar chart. The mental health comprehensive report submodule uses large-scale model-based prompt word engineering technology, based on the conventional mental health comprehensive report structure, combined with multiple mental health analysis reports of college students, to generate composite prompt words, and then the surveyor agent outputs a mental health comprehensive report; The mental health portrait sub-module uses the large model's prompt word engineering technology and multimodal output technology, based on the overview of Plutchik's emotion wheel and the content of conventional mental health animal portraits, combined with the comprehensive mental health report of college students, to combine and generate compound prompt words, and then the surveyor intelligent agent outputs a mental health animal portrait.

4. The AI ​​Agent-based college student mental health service system according to claim 1 is characterized in that: The social personality portrait module includes: The web crawler module obtains the uniform resource locators of the posts through the social media accounts provided by college students, and then performs multimodal web crawling to obtain the social media data of college students; The descriptive statistics submodule uses large-scale model-based prompt word engineering technology and word cloud graph technology, based on the structure of conventional social media descriptive statistics reports and combined with the social media data of college students to generate compound prompt words. The portrait artist agent then outputs multiple descriptive statistics reports. After systematically processing the social media data of college students, the sentiment analysis submodule uses large-scale model-based prompt word engineering technology and SnowNLP sentiment analysis technology. Based on the structure of conventional social media sentiment analysis reports, it combines the social media data of college students to generate compound prompt words. The portrait artist agent then outputs a sentiment analysis report. The MBTI personality analysis submodule leverages the large model's prompt word engineering technology and multimodal output technology, based on the MBTI scale's overview, scoring criteria, and conventional personality analysis report structure, combined with the social media data of college students, to generate composite prompt words. The portrait artist agent then outputs an MBTI personality inference report and an MBTI personality portrait.

5. The AI ​​Agent-based college student mental health service system according to claim 1 is characterized in that: The mental health intervention module includes: The database submodule connects the counselor's intelligent body to the psychological counseling expert guide and real counseling case data. It uses prompt word engineering technology to connect the counselor's intelligent body to the psychological counseling expert guide. It also extracts information through prompt word engineering technology to obtain a keyword list. This keyword list is then vectorized using a vectorized embedding model and stored in a vector library, connecting the counselor's intelligent body to real counseling case data. The adaptive recommendation submodule leverages the large-scale model's prompt word engineering technology, based on the overview of the psychological counseling multidimensional matching model and the structure of conventional counselor recommendation reports. It combines data from the college students' mental health measurement module and social personality profiling module to generate composite prompt words, and then outputs a counselor agent recommendation report. The psychological counseling service sub-module uses the large model's prompt word engineering technology, retrieval enhancement generation technology and thinking chain technology, based on the psychological counseling expert guide and real counseling case data, combined with the college students' mental health measurement module data and social personality portrait module data, to combine and generate compound prompt words, and then the counselor intelligent body provides professional and effective mental health intervention services to college students.

6. The AI ​​Agent-based mental health service method for college students is characterized by: The system according to any one of claims 1 to 5 is implemented, comprising the following steps: S1: The surveyor agent calculates the multidimensional mental health scores of college students using a standardized mental health scale; S2: The surveyor agent leverages the large-scale model's prompt word engineering and multimodal output technology to generate multiple mental health analysis reports, a mental health radar chart, a mental health summary report, and a mental health animal portrait. S3: The portrait agent uses the social media data of college students obtained by web crawlers to generate multiple descriptive statistics reports and a sentiment analysis report; S4: The portrait artist agent uses the large model's prompt word engineering technology and multimodal output technology to generate an MBTI personality prediction report and an MBTI personality portrait; S5: The counselor agent is connected to the psychological counseling expert guide and real counseling case data. Based on the data of the mental health measurement module and the social personality profile module, the system adaptively recommends counselor agents for college students. S6: The counselor intelligent agent uses the large model's prompt word engineering technology, retrieval enhancement generation technology and thinking chain technology to provide professional and effective mental health intervention services to college students.

Images