Method and system for establishing teenager development comprehensive evaluation model

Abstract

The invention discloses a system for generating a personalized intervention scheme based on comprehensive evaluation of teenager development. The system comprises a terminal side used for collecting teenager physiological behavior data, core cognitive ability data, physical quality data and psychological data and receiving the personalized intervention scheme for teenager growth; the cloud platform is used for processing the physiological behavior data, the core cognitive ability data, the physical quality data and the psychological data of the teenagers collected by the terminal side and generating a personalized intervention scheme for the growth of the teenagers according to a processing result; according to the method, layer-by-layer evaluation and prediction from a physiological basis to psychological adjustment to capability expression can be realized through a nonlinear mapping function, and personalized course recommendation and psychological adjustment scheme generation can be realized.

Description

Technical Field

[0001] This invention relates to the field of educational assessment and mental health monitoring technology, specifically to a method and system for establishing a comprehensive assessment model for adolescent development. Background Technology

[0002] Adolescents face multiple challenges during their development, including academic pressure, mental health crises, and internet addiction. Traditional psychological assessments primarily rely on subjective, static scales (such as PHQ-9, GAD-7, MHT, EPQ-C, etc.) and manual observation and feedback. The limitations of current technologies include:

[0003] 1. Highly subjective: It relies on the experience and judgment of the evaluator, and the results are easily affected by subjective bias.

[0004] 2. Limited data: Data is obtained solely through questionnaires or behavioral observations, which cannot comprehensively cover multiple dimensions such as physiological and psychological aspects.

[0005] 3. Lack of dynamism: It is impossible to track the changing trends in the growth process of teenagers in real time.

[0006] 4. Insufficient predictive ability: The assessment of ability development mainly relies on teachers' subjective evaluation and lacks objective quantitative indicators, resulting in a serious lack of early predictive ability for psychological problems (such as learning anxiety and interpersonal anxiety) and potential development.

[0007] With the development of artificial intelligence technology, the application of multimodal data fusion (such as physiological signals and behavior), machine learning, and deep learning models has provided new ideas for adolescent growth assessment. However, existing technologies have not yet formed a systematic, multi-layered, progressive assessment framework, making it difficult to cover the full-chain assessment needs of diverse potentials from the physiological foundation level.

[0008] Therefore, a systematic technical solution is needed that can integrate multimodal data and achieve dynamic assessment and personalized intervention recommendations. Summary of the Invention

[0009] The purpose of this invention is to provide a method and system for establishing a comprehensive assessment model for adolescent development, in order to overcome the shortcomings of existing technologies.

[0010] According to a first aspect of the present invention, a system for generating personalized intervention programs based on a comprehensive assessment of adolescent development includes:

[0011] The terminal is used to collect physiological and behavioral data, core cognitive ability data, physical fitness data, and psychological data of adolescents, and to receive personalized intervention plans for adolescent growth.

[0012] A cloud platform for processing physiological and behavioral data, core cognitive ability data, physical fitness data, and psychological data of adolescents collected from the terminal side, and generating personalized intervention plans for adolescent growth based on the processing results.

[0013] The remote platform includes: a data integration module that integrates the physiological behavior data, core cognitive ability data, physical fitness data, and psychological dataset into a multidimensional assessment dataset for evaluation; a multidimensional assessment model that performs dynamic evaluation using the multidimensional assessment dataset; a prediction module that predicts the psychological health risks and potential development trajectories of adolescents based on the dynamic assessment results output by the multidimensional assessment model; and an intervention strategy generation module that generates and sends the personalized intervention plan to the terminal based on the prediction results output by the prediction module.

[0014] Preferably, the terminal side includes: a physiological behavior data acquisition module for collecting adolescent physiological behavior data, including EEG data and HRV data; a core cognitive ability data acquisition module for collecting adolescent core cognitive ability data, which includes visual perception test data, visual perception test data, auditory perception test data, auditory perception test data, selective attention data, sustained attention data, alternating attention data, distributed attention and reaction attention test data, sensory memory test data, working memory test data, and long-term memory test data; a physical fitness data acquisition module for collecting adolescent physical fitness data, including daily physical fitness test scores from school; a psychological data acquisition module for collecting adolescent psychological data, including mental health and personality traits; and a transceiver module for sending data to a cloud platform and successfully receiving the personalized intervention plan on the cloud platform.

[0015] Preferably, the multidimensional assessment model is a multi-level progressive mapping model established by nonlinear functions g and f, ranging from physiological aspects including brain cognition and physical fitness to psychological aspects including psychological resilience and mental health, and further to practical abilities, social abilities, and multiple potentials.

[0016] Preferably, the multi-level progressive mapping model includes: a physiological foundation layer for assessing core cognitive abilities and physical fitness; a psychological adjustment layer obtained by mapping the physiological foundation layer through a nonlinear function g, wherein the psychological adjustment layer is used to analyze quantitative indicators of psychological resilience, psychological health status, and personality traits; an ability performance layer obtained by mapping the psychological adjustment layer through a nonlinear function f, wherein the ability performance layer is used to predict practical abilities, social abilities, multiple potentials, and personality traits; and a multiple potential layer obtained by mapping the psychological adjustment layer through a nonlinear function f, wherein the multiple potential layer is used to predict the potential of linguistic intelligence, logical-mathematical intelligence, spatial intelligence, bodily-kinesthetic intelligence, musical intelligence, interpersonal intelligence, self-awareness intelligence, and naturalistic intelligence.

[0017] Preferably, the present invention utilizes the Transformer model to realize the function of the nonlinear function g of the psychological adjustment layer.

[0018] Preferably, the physiological basis layer of the present invention assesses core cognitive abilities and physical fitness based on a multidimensional assessment dataset output by the data integration module.

[0019] Preferably, the core cognitive abilities assessed by the physiological basal layer of the present invention include vision, hearing, attention, and memory; the physical qualities assessed by the physiological basal layer include sleep quality, cardiopulmonary function, motor coordination, height, weight, NMI index, and five basic items.

[0020] Preferably, the present invention employs a feature alignment algorithm and PCA dimensionality reduction to perform multimodal data fusion on the brain core cognitive ability, HRV, physical fitness, psychological resilience, mental health, and multiple potential outputs of the multidimensional assessment model, so as to obtain multimodal fused data.

[0021] Preferably, the prediction module predicts mental health risks and potential development trajectories by processing the multimodal fusion data with logistic regression and K-means clustering.

[0022] Preferably, the intervention strategy generation module generates personalized course recommendations and psychological adjustment plans based on reinforcement learning (RL).

[0023] According to a second aspect of the present invention, a method for establishing a comprehensive assessment model for adolescent development includes: a multi-level progressive mapping model established by nonlinear functions g and f, ranging from physiological aspects including brain cognition and physical fitness to psychological aspects including psychological resilience and mental health, and further to practical abilities, social abilities, and multiple potentials.

[0024] Preferably, the multi-level progressive mapping model of the present invention includes: a physiological foundation layer for assessing core cognitive abilities and physical fitness; a psychological adjustment layer obtained by mapping the physiological foundation layer through a nonlinear function g, wherein the psychological adjustment layer is used to analyze quantitative indicators of psychological resilience, quantitative indicators of mental health status, and quantitative indicators of personality traits; an ability performance layer obtained by mapping the psychological adjustment layer through a nonlinear function f, wherein the ability performance layer is used to predict practical abilities, social abilities, multiple potentials, and personality traits; and a multiple potential layer obtained by mapping the psychological adjustment layer through a nonlinear function f, wherein the multiple potential layer is used to predict the potential of linguistic intelligence, logical-mathematical intelligence, spatial intelligence, bodily-kinesthetic intelligence, musical intelligence, interpersonal intelligence, self-awareness intelligence, and naturalistic intelligence.

[0025] The above-described technical solution of the present invention can solve the following technical problems:

[0026] Construct a multi-dimensional, multi-level assessment model that covers core indicators such as physiology, ability, and potential.

[0027] The nonlinear mapping function enables a step-by-step assessment and prediction from physiological basis to psychological adjustment and then to ability performance.

[0028] Based on reinforcement learning, dynamic intervention strategies are designed to achieve personalized course recommendations and psychological adjustment plans.

[0029] It provides highly interpretable data modeling and algorithm design to ensure the scientific validity and practicality of the model.

[0030] Objective psychological assessment, personalized intervention and closed-loop management: By using multimodal data fusion and multi-level progressive mapping assessment models, subjective bias is reduced, assessment accuracy is improved, and a full-process system of "assessment-early warning-intervention-feedback" is formed to improve the efficiency of mental health management. Attached Figure Description

[0031] Figure 1 This is a schematic diagram of the system for generating personalized intervention plans based on a comprehensive assessment of adolescent development according to the present invention;

[0032] Figure 2 This is a schematic diagram of the multi-level progressive mapping model of the present invention;

[0033] Figure 3 This is a schematic diagram of the multi-level progressive mapping relationship of the present invention. Detailed Implementation

[0034] This invention utilizes a multi-level progressive mapping model based on multimodal data fusion, dynamic assessment, and prediction to quantitatively assess the growth status, psychological resilience, ability performance, and potential development of adolescents. By constructing a multi-level progressive mapping model from physiological to psychological to ability-based approaches, and through multimodal data collection, dynamic assessment, and prediction, this invention achieves comprehensive quantitative analysis of adolescent growth status and personalized intervention management. The multi-level progressive mapping model specifically includes:

[0035] Physiological foundation layer (brain cognition and physical development conditions): assess brain cognitive abilities (4 indicators: vision, hearing, attention, memory, etc.) and body values ​​(5 indicators: height and weight, BMI (body mass index), motor coordination, cardiopulmonary function, sleep quality, etc.).

[0036] Psychological adjustment layer (conditions for psychological development): Analyzes psychological resilience (stress recovery ability, efficiency of emotion regulation), personality traits (psychoticism, neuroticism, introversion / extroversion), and mental health status (learning anxiety, interpersonal anxiety, depressive tendencies, etc.).

[0037] The ability performance layer (conditions for ability development): quantifies the realization of abilities (creativity, problem-solving ability) and social abilities (teamwork, leadership).

[0038] Multiple potential layer: predicts the development potential of eight categories of potential, including linguistic intelligence, logical-mathematical intelligence, spatial intelligence, bodily-kinesthetic intelligence, musical intelligence, interpersonal intelligence, intrapersonal intelligence, and nature exploration.

[0039] The system comprises three core modules: data acquisition and preprocessing, multidimensional assessment, and prediction and intervention recommendations. First, through data acquisition, it integrates multidimensional data on adolescents' physiological and psychological performance to construct a multidimensional assessment dataset. Data preprocessing is then performed, including feature extraction, standardization, and fusion. Second, an AI model is used to conduct a stratified assessment of adolescents' developmental status, ensuring accuracy. Finally, personalized intervention plans are generated based on the assessment results. This achieves a comprehensive assessment of adolescents' positive development and health. This method has significant application value in the field of adolescent psychological and physiological health assessment technology.

[0040] Specifically, the present invention provides a system for generating personalized intervention plans based on a comprehensive assessment of adolescent development, such as... Figure 1 As shown, it includes:

[0041] Terminal 100 is used to collect physiological and behavioral data, core cognitive ability data, physical fitness data, and psychological data of adolescents and to receive personalized intervention plans for adolescent growth.

[0042] A cloud platform 200 is used to process the physiological and behavioral data, core cognitive ability data, physical fitness data and psychological data of adolescents collected by the terminal side 100 and generate the personalized intervention plan for adolescent growth based on the processing results.

[0043] The remote platform 200 includes: a data integration module 210 that integrates the physiological behavior data, core cognitive ability data, physical fitness data, and psychological dataset into a multidimensional assessment dataset for evaluation; a multidimensional assessment model 220 that performs dynamic evaluation using the multidimensional assessment dataset; a prediction module 230 that predicts the psychological health risks and potential development trajectory of adolescents based on the dynamic assessment results output by the multidimensional assessment model 220; and an intervention strategy generation module 240 that generates and sends the personalized intervention plan to the terminal side 100 based on the prediction results output by the prediction module 230.

[0044] Figure 1The terminal side 100 shown includes: a physiological behavior data acquisition module 110 for collecting adolescent physiological behavior data, including EEG data and HRV data; a core cognitive ability data acquisition module 120 for collecting adolescent core cognitive ability data, which includes visual perception test data, visual perception test data, auditory perception test data, auditory perception test data, selective attention data, sustained attention data, alternating attention data, distributed attention and reaction attention test data, sensory memory test data, working memory test data, and long-term memory test data; a physical fitness data acquisition module 130 for collecting adolescent physical fitness data, including daily physical fitness test scores from school; a psychological data acquisition module 140 for collecting adolescent psychological data, including mental health and personality traits; and a transceiver module for sending data to the cloud platform 200 and successfully receiving the personalized intervention plan on the cloud platform 200.

[0045] Figure 1 The multidimensional assessment model 220 shown is a multi-level progressive mapping model established by nonlinear functions g and f, ranging from physiological aspects including brain cognition and physical fitness to psychological aspects including psychological resilience and mental health, and further to practical abilities, social abilities, and multiple potentials.

[0046] See Figure 2 and Figure 3 The multi-level progressive mapping model of the present invention includes: a physiological foundation layer for assessing core cognitive abilities and physical fitness; a psychological adjustment layer obtained by mapping the physiological foundation layer through a nonlinear function g, wherein the psychological adjustment layer is used to analyze quantitative indicators of psychological resilience, quantitative indicators of mental health status, and quantitative indicators of personality traits; an ability performance layer obtained by mapping the psychological adjustment layer through a nonlinear function f, wherein the ability performance layer is used to predict practical abilities, social abilities, multiple potentials, and personality traits; and a multiple potential layer obtained by mapping the psychological adjustment layer through a nonlinear function f, wherein the multiple potential layer is used to predict the potential of linguistic intelligence, logical-mathematical intelligence, spatial intelligence, bodily-kinesthetic intelligence, musical intelligence, interpersonal intelligence, self-awareness intelligence, and naturalistic intelligence.

[0047] This invention utilizes the Transformer model to realize the function of the nonlinear function g in the psychological adjustment layer.

[0048] The physiological basis layer of this invention assesses core cognitive abilities and physical fitness based on the multidimensional assessment dataset output by the data integration module 210.

[0049] The core cognitive abilities assessed by the physiological basal layer of this invention include vision, hearing, attention, and memory; the physical qualities assessed by the physiological basal layer include sleep quality, cardiopulmonary function, motor coordination, height, weight, NMI index, and five basic parameters.

[0050] This invention employs a feature alignment algorithm and PCA dimensionality reduction to perform multimodal data fusion on the brain core cognitive ability, HRV, physical fitness, psychological resilience, mental health, and multiple potential outputs of the multidimensional assessment model 220, in order to obtain multimodal fused data.

[0051] Figure 1 The prediction module 230 shown predicts mental health risks and potential development trajectories by processing the multimodal fusion data with logistic regression and K-means clustering.

[0052] Figure 1 The intervention strategy generation module 240 shown generates personalized course recommendations and psychological adjustment plans based on reinforcement learning (RL).

[0053] See Figure 2 and Figure 3 The present invention provides a method for establishing a comprehensive assessment model for adolescent development, which includes: establishing a multi-level progressive mapping model through nonlinear functions g and f, from physiological aspects including brain cognition and physical fitness to psychological aspects including psychological resilience and mental health, and further to practical abilities, social abilities, and multiple potentials.

[0054] The multi-level progressive mapping model of this invention includes: a physiological foundation layer for assessing core cognitive abilities and physical fitness; a psychological adjustment layer obtained by mapping the physiological foundation layer through a nonlinear function g, wherein the psychological adjustment layer is used to analyze quantitative indicators of psychological resilience, psychological health status, and personality traits; an ability performance layer obtained by mapping the psychological adjustment layer through a nonlinear function f, wherein the ability performance layer is used to predict practical abilities, social abilities, multiple potentials, and personality traits; and a multiple potential layer obtained by mapping the psychological adjustment layer through a nonlinear function f, wherein the multiple potential layer is used to predict the potential of linguistic intelligence, logical-mathematical intelligence, spatial intelligence, bodily-kinesthetic intelligence, musical intelligence, interpersonal intelligence, self-awareness intelligence, and naturalistic intelligence.

[0055] The technical solution of the present invention will be described in detail below through specific examples.

[0056] This invention relates to a multi-level progressive mapping evaluation model and system based on AI-guided assessment and multimodal data fusion, comprising the following core components:

[0057] ● Multi-level progressive mapping model: A dynamic assessment framework is established through nonlinear functions g and f, ranging from physiological (brain cognition, physical fitness) to psychological (psychological resilience, mental health), and then to ability (practical ability, social ability) and potential (8 potential directions).

[0058] ● Multimodal data fusion: Employing feature alignment algorithms and PCA dimensionality reduction, it integrates heterogeneous data such as (core cognitive abilities, HRV, physical fitness, psychological resilience, mental health, and multiple potentials).

[0059] ●System core algorithm design:

[0060] Transformer model: In the psychological regulation layer, it captures long-term dependency through self-attention mechanism and explores the causal chain of "sleep quality → cognitive ability → psychological resilience".

[0061] Prediction submodule: Identifies mental health risks and potential development trajectories through logistic regression and K-means clustering;

[0062] Intervention Submodule: Optimizing personalized course recommendations and psychological adjustment pathways based on reinforcement learning (RL).

[0063] Implementation steps

[0064] The technical solution of this invention aims to construct a comprehensive assessment model and system for adolescent growth status, covering the entire process from data collection to the generation of final intervention recommendations. The system implementation steps are: data collection and preprocessing module, model training and optimization, and system deployment. Specific steps are as follows:

[0065] Data collection:

[0066] Physiological and behavioral data: Real-time collection via wearable devices (such as mood-monitoring smartwatches) includes HRV (Heart Rate Variability) and sleep quality data. Wearable devices monitor minute fluctuations in heart rate intervals (such as changes in the RR interval) to reflect the activity state of the autonomic nervous system. Low HRV values ​​indicate sympathetic nervous system activity (stress / anxiety), while high values ​​indicate parasympathetic nervous system activity (relaxation / pleasure). This data, combined with heart rate, activity level, and sleep quality data, comprehensively assesses emotional state, ultimately providing quantitative data across five dimensions: stress, anxiety, pleasure, relaxation, and sleep quality.

[0067] Core cognitive ability data: obtained through interactive cognitive behavioral tests. These include 12 indicators such as visual perception test, visual perception test, auditory perception test, auditory perception test, selective attention, sustained attention, alternating attention, distributed attention and reaction attention test, sensory memory test, working memory test and long-term memory test.

[0068] Physical fitness data: Import or input through daily physical test scores from the school (such as 50-meter run, rope skipping, lung capacity, etc.).

[0069] Psychological data: Combining labeled assessment scales, such as MHT (Mental Health), EPQ-C (Personality Traits), CD-RISC scale (Psychological Resilience), and AI-generated interactive game data.

[0070] Data preprocessing

[0071] Feature alignment: Using adversarial learning to eliminate distributional differences in multimodal data;

[0072] PCA dimensionality reduction: retains more than 95% of the information while reducing computational complexity.

[0073] Model training and optimization

[0074] Multi-level progressive mapping:

[0075] ●The mapping from physiological to psychological states (g):

[0076] Ψ=g(CCA,PF)=σ(W g ·[CCA,PF]+b g )

[0077] The concatenated input vector;

[0078] σ(●): ReLU activation function;

[0079] W g ∈R 17×20 ,b g ∈R 17 Learnable parameters.

[0080] ●The mapping of psychology to ability (f):

[0081] Φ=f(Ψ)=σ(W f ·Ψ+b f )

[0082] W f ∈R 16×17 ,b f ∈R 16 Learnable parameters.

[0083] Transformer modeling: In the psychological regulation layer, long-term dependencies (such as periodic changes in anxiety symptoms or causal chains from sleep quality → cognitive function → psychological resilience) are captured through self-attention mechanisms.

[0084] Reinforcement learning intervention: Defining the state space S t(Assessment results, training completion rate), motion space A t (Training combination, psychological adjustment strategy) optimizes the intervention path by maximizing long-term rewards.

[0085] Mathematical Modeling and Algorithms

[0086] Model building methods

[0087] Construct a multi-dimensional, multi-level progressive mapping model covering core indicators such as physiology, ability, and potential. Specifically, this includes:

[0088] Physiological Foundation Layer (Brain Cognition and Physical Development Conditions): This layer assesses core cognitive abilities and physical fitness. Core cognitive abilities primarily include 12 quantifiable indicators: vision (visual sensation, visual perception), hearing (auditory sensation, auditory perception), attention (selective attention, sustained attention, alternating attention, distributed attention, and reactive attention), and memory (sensory memory, working memory, and long-term memory). Physical fitness includes sleep quality, cardiopulmonary function, motor coordination, height, weight, BMI, and five basic components (strength, speed, endurance, agility, and flexibility), totaling 11 quantifiable indicators.

[0089] Psychological adjustment layer (psychological development conditions): Analyzes 5 quantitative indicators of psychological resilience (including stress resilience, frustration tolerance, emotion regulation efficiency, cognitive flexibility, and self-acceptance), 9 quantitative indicators of mental health status (learning anxiety, interpersonal anxiety, loneliness tendency, self-blame tendency, hypersensitivity tendency, somatization tendency, phobic tendency, impulsivity tendency, and depressive tendency), and 3 quantitative indicators of personality traits (psychoticism, neuroticism, and introversion / extroversion).

[0090] The competency performance layer (conditions for growth and development) includes a total of 8 quantitative indicators, including practical abilities (creativity, problem-solving, resource integration, trend prediction, technology application, etc.) and social abilities (teamwork, conflict resolution, leadership).

[0091] Multiple potential layers (potential development conditions): predicting potential in eight areas, including linguistic intelligence, logical-mathematical intelligence, spatial intelligence, bodily-kinesthetic intelligence, musical intelligence, interpersonal intelligence, intrapersonal intelligence, and naturalistic intelligence.

[0092] Mathematical expression and modeling of the model

[0093] Mathematical expression of the model

[0094] ● Physiological Base Layer

[0095] Core Cognitive Abilities (CCA)

[0096] ◆Visual Perception (VP)

[0097] VP={V p V s},in:

[0098] V p Visual Perception

[0099] V s Visual Sensation

[0100] ◆Auditory Perception (AP)

[0101] AP = {A p A s},in:

[0102] A p Auditory Perception

[0103] A s Auditory Sensation

[0104] ◆Attention (ATT)

[0105] ATT = {ATT s ,ATT c ,ATT a ,ATT d ,ATT r},in:

[0106] ATT s Selective Attention

[0107] ATT c Sustained attention

[0108] ATT a Alternating Attention

[0109] ATT d Divided Attention

[0110] ATT r Response Attention

[0111] ◆Memory (MEM)

[0112] MEM = {MEM s MEM w MEM l},in:

[0113] MEM s Sensory memory

[0114] MEM w Working memory

[0115] MEM l Long-term memory

[0116] Total core cognitive ability vector:

[0117]

[0118] Physical Fitness (PF)

[0119] ◆Basic Physical Attributes (BPA)

[0120] BPA = {F, S, E, A, F}, where:

[0121] F: Strength

[0122] S: Speed

[0123] E: Endurance

[0124] A: Agility

[0125] F: Flexibility

[0126] ◆Additional Physical Attributes (APA)

[0127] APA = {Sleep, HeartLung, Coordination}, where:

[0128] Sleep: Sleep Quality

[0129] HeartLung: Cardiopulmonary function (cardiopulmonary fitness)

[0130] Coordination: Motor Coordination

[0131] Total physical fitness vector:

[0132]

[0133] ● Psychological Regulation Layer

[0134] Psychological Resilience (PR)

[0135] in:

[0136] R stress Stress Recovery

[0137] R frustration Frustration Tolerance

[0138] R emotion Emotional Regulation Efficiency

[0139] R cognition Cognitive Flexibility

[0140] R self Self-acceptance

[0141] Mental Health (MH)

[0142]

[0143] ,in:

[0144] Anxiety study Study Anxiety

[0145] Anxiety social Social Anxiety

[0146] Loneliness: a tendency towards loneliness

[0147] Guilt: Guilt Tendency

[0148] Allergy: Hypersensitivity Tendency

[0149] Somatic: Somatic Symptoms

[0150] Phobia: Phobia Tendency

[0151] Impulsivity: Impulsivity Tendency

[0152] Depression: Depression Tendency

[0153] Personality Traits (PT)

[0154] in:

[0155] P: Psychoticism

[0156] N: Neuroticism

[0157] E: Extraversion / Introversion

[0158] Comprehensive vector of psychological adjustment layer:

[0159]

[0160] ●Competence Manifestation Layer

[0161] Practical competence (PC)

[0162]

[0163] ,in:

[0164] Creativity: Creativity

[0165] Problem Solving Ability

[0166] Resource Integration: Resource integration capability

[0167] RiskPrediction: The ability to predict and anticipate risks.

[0168] TechApplication: Technical Application Capability

[0169] Social Competence (SC)

[0170] in:

[0171] Teamwork: Teamwork Ability

[0172] ConflictResolution: Conflict Resolution Ability

[0173] Leadership: Leadership Skills

[0174] ●Multi-potential layer

[0175] Multiple Potentials (MP)

[0176] in:

[0177] L: Linguistic Intelligence

[0178] M: Logical-Mathematical Intelligence

[0179] S: Spatial Intelligence

[0180] B: Bodily-Kinesthetic Intelligence

[0181] Mu: Musical Intelligence

[0182] I: Interpersonal Intelligence

[0183] S C Intrapersonal Intelligence

[0184] N C Naturalistic Intelligence

[0185] Comprehensive performance vector (including potential):

[0186]

[0187] Mathematical modeling

[0188] ●Multi-level progressive mapping function:

[0189] enter:

[0190] First layer: Physiological Foundation (PF)

[0191] ●Includes:

[0192] ■ Core Cognitive Ability (CCA): 12 indicators (such as vision, hearing, attention, memory, etc.)

[0193] ■ Physical Fitness (PF): 8 indicators (such as strength, endurance, etc.).

[0194] ●Total dimensions: 12 + 8 = 20-dimensional vector

[0195] Notation:

[0196]

[0197] The second layer: Psychological Regulation (Ψ)

[0198] ●Includes:

[0199] ■ Psychological Resilience: 5 indicators;

[0200] ■Mental Health: 9 indicators;

[0201] ■ Three indicators of personality traits;

[0202] ●Total dimensions: 5 + 9 + 3 = 17-dimensional vector

[0203] Notation:

[0204] Ψ={Resilience,MH,PT}∈R 17

[0205] Output:

[0206] The third and fourth layers: Competency Performance (Φ) and Multi-faceted Potential Layer.

[0207] ●Includes:

[0208] ■ Practical Competence: 5 indicators;

[0209] ■ Social Competence: 3 indicators;

[0210] ●Total dimensions: 5 + 3 = 8-dimensional vector

[0211] ●Add multiple potential prediction (8 types of intelligence): 8+8=16-dimensional vector

[0212] Notation:

[0213] Φ={PC,SC,MP}∈R 16

[0214] The above four layers are integrated into a progressive mathematical model, defined as a functional relationship:

[0215] Φ=f(Ψ)=f(g(CCA,PF)), where:

[0216] CCA∈R 12 ,PF∈R 18

[0217] Ψ=g(CCA,PF)∈R 17

[0218] Φ=F(Ψ)∈R 16

[0219] Nonlinear mapping:

[0220] ●The mapping from physiological to psychological aspects (g)

[0221] function g:R 20 →R 17It is a mapping function from the physiological foundation layer to the psychological regulation layer, representing the supporting role of physiological and cognitive abilities on psychological state, and its form is as follows:

[0222] Ψ=g(CCA,PF)=W g ·[CCA;PF]+b g ,in:

[0223] [CCA;PF] represents a concatenation operation that produces a 20-dimensional input vector;

[0224] W g ∈R 17×20 It is a weight matrix;

[0225] b g ∈R 17 It is a bias term.

[0226] Its optional nonlinear extensions:

[0227] Ψ=σ(W g ·[CCA;PF]+b g )

[0228] Where σ is the activation function (such as ReLU), used to introduce nonlinear relationships.

[0229] This function represents the "supporting role of physiological and cognitive abilities on psychological state." It can be understood as follows: the better an adolescent's physical fitness and core cognitive abilities, the easier it is for them to develop good psychological resilience, emotional regulation abilities, and personality traits.

[0230] ● The mapping from psychology to ability (f)

[0231] function f:R 17 →R 16 This is a mapping function from the psychological adjustment layer to the ability performance layer, representing the driving role of psychological qualities in ability development. Its form is as follows:

[0232] Φ=f(Ψ)=W f ·Ψ+b f ,in:

[0233] W f ∈R 16×17 It is a weight matrix;

[0234] b f ∈R 16 It is a bias term;

[0235] Its optional nonlinear extensions:

[0236] Φ=σ(W f ·Ψ+b f )

[0237] Where σ is a nonlinear activation function (such as ReLU).

[0238] This function expresses the "driving effect of psychological qualities on ability development": it can be understood that the better the psychological resilience, mental health level and personality characteristics of adolescents, the more they can promote the development of practical and social abilities and stimulate multiple potentials.

[0239] In summary, the entire multi-layered progressive model can be represented as:

[0240] in

[0241] CCA∈R 12 ,PF∈R 8 ,Ψ∈R 17 ,Φ∈R 16

[0242] σ(·) is a nonlinear activation function (such as ReLU).

[0243] This model intuitively reflects the progressive logic from physiological basis to psychological adjustment and then to ability performance through nested hierarchical functions, and quantifies the dynamic relationship between each level through weight coefficients and nonlinear activation functions.

[0244] ● Multimodal data fusion

[0245] Adversarial learning feature alignment:

[0246] in

[0247] L task : Psychological state classification loss.

[0248] L domain Domain-specific loss determination.

[0249] PCA dimensionality reduction: retains 95% of the information and reduces 20-dimensional features to 10-dimensional features.

[0250] Core Algorithm Design

[0251] The core algorithm design of this invention includes the application of the SUM classifier (classifying psychological resilience levels), the Transformer model (capturing long-term dependent causal chains), the logistic regression algorithm (predicting mental health risks), the K-means clustering algorithm (predicting potential development trajectories), and reinforcement learning (RL) (personalized training course recommendations). These components work together to extract information on the psychological state of adolescents from multimodal data, assess psychological resilience and mental health, predict mental health status and potential, and promote their positive development by dynamically adjusting intervention strategies.

[0252] Applications of SVM classifiers

[0253] The Support Vector Machine (SVM) classifier is a classic classification algorithm. Its core idea is to find an "optimal hyperplane" that separates data of different categories as much as possible. In this invention, an SVM classifier is used to classify data into categories with high and low psychological resilience. Specifically:

[0254] Input: Score of psychological assessment scale (CD-RISC psychological resilience assessment scale score)

[0255] Output: Classification results (high psychological resilience or low psychological resilience)

[0256] Implementation steps

[0257] Data collection and preprocessing

[0258] Data source: Psychological data of participants were collected using the CD-RISC (Connor-Davidson Psychological Resilience Scale) and the PHQ-9 (Patient Health Questionnaire-9) psychological assessment scale.

[0259] Feature engineering: Transforming raw scores into numerical features (such as total score, dimensional mean score).

[0260] Standardization: Normalize the data (Z-score standardization) to ensure that the characteristics of different scales are comparable.

[0261] Model training:

[0262] Split the dataset: Divide the dataset into a training set and a test set (80% training, 20% test).

[0263] Kernel function selection: The RBF kernel (Radial Basis Function kernel) is suitable for handling nonlinear relationships, therefore the RBF kernel is chosen as the kernel function for the model. The kernel function determines how the data is mapped to a high-dimensional space.

[0264] Training the model: The optimal plane is found through the SMO (Sequence Minimum Optimization) algorithm to minimize the classification error.

[0265] Interpretation of results:

[0266] Classification accuracy: Evaluates the model's performance on the test set, such as accuracy and F1 score.

[0267] Feature importance: Analyze which scale questions have a greater impact on the classification results by weighting coefficients. If the data is unbalanced (e.g., low psychological resilience results with few samples), the weights need to be adjusted.

[0268] Transformer model: capturing long-term dependent causal chains

[0269] When analyzing changes in adolescents' psychological states, we used the Transformer model, which can handle complex multimodal time series data and effectively extract long-term and short-term dependencies. For example, it can capture the periodic changes in anxiety symptoms and explore causal chain modeling of "sleep quality → cognitive function → psychological resilience".

[0270] Input Representation

[0271] The input data is time-series data, with each time point containing multiple features (such as anxiety score, heart rate variability, sleep quality, cognitive ability score, etc.). Let the time length be T, and each feature dimension be D, then the input matrix is ​​X∈R. {T×D} .

[0272] Self-attention mechanism

[0273] Self-attention allows the model to focus on the relationships between different time points. The calculation formula is as follows:

[0274]

[0275] in:

[0276] Q, K, and V are the query, key, and value matrices, respectively.

[0277] d k This is a scaling factor to prevent the gradient from vanishing due to an excessively large inner product.

[0278] This mechanism allows the model to focus on key historical nodes that influence it when processing information at the current moment, thereby more accurately modeling the evolution of mental states.

[0279] For example, suppose we have daily anxiety scores and sleep quality data from the past week. Using self-attention mechanisms, we can find out which day's anxiety score affected the overall trend.

[0280] Feed-Forward Network

[0281] Each attention layer is typically followed by a feedforward neural network (FFN) to further process the extracted information. For example, an FFN can help the model better understand how sleep deprivation affects cognitive performance the next day, and consequently, long-term psychological resilience.

[0282] Causal chain analysis of "sleep quality → cognitive function → psychological resilience"

[0283] Using the Transformer model, we can analyze the causal chain from sleep quality to cognitive function and then to psychological resilience. The specific steps are as follows:

[0284] ● Input: Multimodal data such as daily anxiety score, sleep quality, and heart rate variability;

[0285] ●Output: Predicted psychological resilience score for a future period of time;

[0286] ●Model training: The model is trained using a large amount of historical data to accurately predict and explain the causal relationships between various factors.

[0287] Prediction Sub-module 1: Logistic Regression – Identification of Mental Health Risks

[0288] This invention employs a logistic regression (LR) model as a predictive submodule for identifying mental health risks. This model offers significant advantages in interpretability, training efficiency, and deployment cost. The goal of this predictive submodule is to identify student groups with potential mental health risks (such as learning anxiety or depressive symptoms) to support early intervention.

[0289] 1. Model Building

[0290] (1) Mathematical expression

[0291] The basic form of the logistic regression model is as follows:

[0292] P(Y=1|X=σ(W) T X+b)

[0293] in:

[0294] Y∈{0,1} represents the target variable, and the mental health risk label (1: at risk; 0: no risk); X∈R d It is the input feature vector, which contains multiple variables related to mental health status;

[0295] W∈R d These are the model's weight parameters, representing the degree of influence each feature has on the output result;

[0296] b∈R: The bias term is used to adjust the offset of the overall output of the model;

[0297] It is the Sigmoid function, which maps the linear output to the interval [0, 1], representing the probability of the positive class (having mental health risks).

[0298] ●Input features:

[0299] The input feature X comes from multi-source data fusion, including but not limited to the following dimensions:

[0300]

[0301] ●Output:

[0302] Binary classification results: Is there a mental health risk? (1: Yes; 0: No)

[0303] (2) Model training process

[0304] ●Data Preprocessing

[0305] Missing value imputation (such as mean, median, or interpolation).

[0306] Feature normalization (Z-Score standardization)

[0307] Category is also being encoded (One-Hot Encoding)

[0308] ●Label Definition

[0309] Define mental health risk labels Y using professional psychological scales (such as PHQ-90, GAD-7) or expert-annotated data.

[0310] ●Loss Function

[0311] Optimization is performed using the cross-entropy loss function:

[0312]

[0313] in:

[0314] N: Total number of samples;

[0315] y i : The true label of the i-th sample;

[0316] The probability predicted by the model;

[0317] ●Optimization Methods

[0318] Minimize the loss function using gradient descent (SGD / Adam);

[0319] L1 / L2 regularization can be added to prevent fitting.

[0320] ●Model Output

[0321] The output is the probability of mental health risk for each individual;

[0322] Set a threshold (e.g., 0.5) for classification decisions.

[0323] Model evaluation metrics

[0324] To comprehensively evaluate the performance of logistic regression models in mental health risk identification tasks, this invention employs four core evaluation metrics:

[0325] (1) Accuracy

[0326] It measures the proportion of samples that the model correctly predicts out of the total sample.

[0327]

[0328] (2) Recall / TPR

[0329] Recall rate is particularly important in the identification of mental health risks.

[0330]

[0331] (3) F1 score

[0332] The harmonic mean of precision and recall is an important metric for evaluating the classification performance of imbalanced datasets.

[0333]

[0334] in:

[0335]

[0336] (4) AUC-ROC curve

[0337] The overall discriminative ability of the model is evaluated by plotting the relationship curve (ROC curve) between the true positive rate (TPR) and the false positive rate (FPR) at different classification thresholds and calculating the area under the curve (AUC).

[0338] AUC=1: Perfect classifier;

[0339] AUC > 0.8: The model performs well;

[0340] AUC≈0.5: The model has no discriminative power.

[0341] Prediction Sub-module 2: K-means Clustering – Identification of Potential Development Trajectory

[0342] This invention employs the K-means clustering algorithm to group adolescents and constructs a potential development trajectory prediction submodule, aiming to identify different potential development patterns or growth trajectories. This not only aids in personalized education path planning but also provides a scientific basis for the precise allocation of educational resources and differentiated training programs.

[0343] 1. Input Features

[0344] To comprehensively reflect the potential development of adolescents, the input features cover data from multiple dimensions:

[0345]

[0346] These features together form a high-dimensional vector XiXi, which describes each student's performance in different dimensions.

[0347] Clustering process

[0348] (1) Data standardization processing

[0349] Because the features have different dimensions and scales, directly using the raw data may lead to some features having an excessive influence on the clustering results. Therefore, data standardization is required before applying K-means. This invention uniformly adopts Z-Score standardization.

[0350] Z-Score standardization: Where μ is the sample mean and σ is the standard deviation.

[0351] (2) Determine the optimal number of clusters k using the elbow method

[0352] Choosing the appropriate number of clusters k is one of the key steps in K-means clustering. The elbow method is a commonly used heuristic approach. Its basic idea is to find the "elbow" as the optimal number of clusters by plotting the relationship between the Within-Cluster Sum of Squares (WCSS) and k for different k values. WCSS is defined as:

[0353]

[0354] in:

[0355] C i Indicates the i-th cluster;

[0356] μ i It is the centroid of the i-th cluster;

[0357] x j It is a sample point belonging to the i-th cluster;

[0358] ||x j -μ i || represents the sample point x j to the center of mass μ i The Euclidean distance.

[0359] By plotting a k vs WCSS graph, we can find the inflection point (i.e., the "elbow point"). The optimal k value corresponding to this point is the optimal number of clusters.

[0360] The specific steps are as follows:

[0361] ● Run the K-means algorithm for different values ​​of k (usually from 1 to a certain upper limit);

[0362] ● Calculate the WCSS value corresponding to each k value;

[0363] ● Plot a k vs WCSS graph to find the inflection point (i.e., the "elbow point") and the optimal k value corresponding to that point.

[0364] This is the optimal number of clusters.

[0365] Clustering samples using the K-means algorithm

[0366] Once the optimal number of clusters k is determined, the K-means algorithm can be formally applied for clustering. The goal of the K-means algorithm is to minimize the sum of squared errors within all clusters. Specifically, given a dataset X = {x1, x2, ..., x...} n Given the number of clusters k, the K-means algorithm attempts to find a set of centroids {μ1, μ2, ..., μ...}. k Minimize the objective function:

[0367]

[0368] The core steps of K-means are as follows:

[0369] ● Initialization: Randomly select k initial centroids {μ1, μ2, ..., μ...} k};

[0370] ● Cluster Assignment: Assign each sample to the nearest centroid based on Euclidean distance. For each sample x j Assign it to the cluster C corresponding to the nearest centroid. i ,Right now:

[0371]

[0372] ●Update centroid: Recalculate the centroid position for each cluster: |C i | represents cluster C i The number of samples in the middle.

[0373] ● Repeated iterations: Repeat steps 2 and 3 until the centroid no longer changes or the maximum number of iterations is reached.

[0374] (3) Visualize and explain each cluster.

[0375] To better understand and interpret clustering results, the following methods can be used:

[0376] ● Two-dimensional / three-dimensional visualization: Dimensionality reduction techniques (such as PCA, t-SNE) are used to project high-dimensional data into a low-dimensional space, and scatter plots are used to show the distribution of various clusters;

[0377] ●Feature analysis: Calculate the average value or distribution of each cluster on each feature, and summarize the typical features;

[0378] ● Tag naming: Based on the feature analysis results, each cluster is given an intuitive and easy-to-understand name, such as "high practice, low social" or "comprehensive and balanced".

[0379] Output

[0380] (1) The developmental trajectory category to which each adolescent belongs

[0381] ● Output the specific cluster number assigned to each student (e.g., "Category 1", "Category 2", etc.). This information can be used for subsequent personalized educational guidance.

[0382] (2) Description of potential development characteristics for each category

[0383] For each cluster, a detailed description of potential development characteristics is provided to help educators understand the common characteristics and developmental trends of different types of students. For example:

[0384] ● High practical skills but low social skills: These students excel in practical abilities and hands-on skills, but are relatively weak in social interaction and teamwork.

[0385] ● Well-rounded type: Possesses well-rounded abilities in all aspects, excels in both theoretical learning and practical operation, and also has good interpersonal skills;

[0386] ●Academic Strength Type: Demonstrates outstanding performance in logical-mathematical and linguistic intelligence, but may lack practical experience and participation in social activities.

[0387] Reinforcement Learning Module: Recommended Personalized Training Courses

[0388] This study utilizes reinforcement learning to optimize intervention strategies, helping adolescents improve their psychological resilience and mental well-being. The goal is to enhance students' practical skills and psychological resilience by recommending personalized training courses or activities.

[0389] Implementation steps:

[0390] Define the state space.

[0391] State space S t Including the current capability layer variable Φt History training course completion rate C t and interest tags I t :

[0392] S t =[Φ t C t ,I t ]∈R 8+5+8

[0393] Φ t : Variables representing abilities, such as creativity score, problem-solving ability score, etc. (8 dimensions);

[0394] C t Recent training course completion status (5 dimensions);

[0395] I t Interest preferences are inferred based on test results, such as scores for linguistic intelligence, logical-mathematical intelligence, musical intelligence, and bodily-kinesthetic intelligence (8 dimensions).

[0396] Define the Action Space

[0397] Action Space A t It includes all possible intervention or course recommendations. For example, an action could be recommending participation in a team project or a self-awareness training session.

[0398] Reward Function

[0399] The reward function is used to measure the effect of taking a certain action:

[0400] R t =α·ΔΦ t +β·CompletionRate t

[0401] in:

[0402] α = 0.6, β = 0.4 indicates the importance of improving balance ability and completing training courses.

[0403] ΔΦ t It refers to the changes in performance variables after taking action;

[0404] Completion Rate t It is the training course completion rate;

[0405] For example, if a course on stress management is recommended, students not only complete the course (increasing the completion rate), but also demonstrate greater psychological resilience in subsequent psychological tests, i.e., ΔΦt If the recommendation is improved, then this recommendation will receive a higher reward score.

[0406] Policy optimization

[0407] Through continuous trial and error and learning, reinforcement learning algorithms can find the optimal intervention strategy and maximize cumulative rewards. This means that the system can dynamically adjust recommended content based on each student's characteristics, achieving truly personalized education. The specific steps are as follows:

[0408] ●Initialization strategy: Randomly select several course recommendation schemes;

[0409] ● Collect feedback: Observe students' performance and record course completion rate and ability improvement;

[0410] ● Update strategy: Adjust the recommendation strategy according to the reward function to gradually approach the optimal solution.

[0411] By applying the Transformer model and reinforcement learning algorithms, this invention enables a deep understanding of the changing patterns of adolescents' psychological states and provides effective interventions to promote their all-round development. The Transformer model helps us capture complex time-dependent and multimodal information, while reinforcement learning allows us to formulate the most suitable growth path based on individual differences. The combination of these two provides strong support for adolescent mental health management and capacity building.

[0412] Technical effects of the present invention

[0413] Improved accuracy of psychological assessments:

[0414] The accuracy of psychological resilience prediction has been improved by 20%.

[0415] The accuracy rate of depression and anxiety screening reached 98%, far exceeding that of traditional scales.

[0416] Enhanced intervention efficiency:

[0417] Training course completion rate increased by 30%, and psychological crisis intervention response time was shortened by 50%.

[0418] Computational efficiency optimization:

[0419] PCA dimensionality reduction reduces computational resource consumption by 50%.

[0420] The Transformer model achieves a 30% improvement in convergence speed during training with 10,000 samples.

[0421] Although the present invention has been described in detail above, it is not limited thereto, and those skilled in the art can make various modifications based on the principles of the present invention. Therefore, all modifications made in accordance with the principles of the present invention should be understood to fall within the protection scope of the present invention.

Claims

1. A system for generating personalized intervention programs based on a comprehensive assessment of adolescent development, comprising: Terminal (100) used to collect physiological and behavioral data, core cognitive ability data, physical fitness data and psychological data of adolescents and to receive personalized intervention plans for adolescent growth. A cloud platform (200) for processing physiological and behavioral data, core cognitive ability data, physical fitness data and psychological data of adolescents collected by the terminal side (100) and generating the personalized intervention plan for adolescent growth based on the processing results; The remote platform (200) includes: The physiological behavior data, core cognitive ability data, physical fitness data and psychological dataset are integrated into a data integration module (210) for a multidimensional assessment dataset used for evaluation; A multidimensional evaluation model (220) for dynamic evaluation using the aforementioned multidimensional evaluation dataset; A prediction module (230) that predicts the mental health risks and potential development trajectory of adolescents based on the dynamic assessment results output by the multidimensional assessment model (220); An intervention strategy generation module (240) generates an intervention strategy based on the prediction results output by the prediction module (230) and sends the personalized intervention plan to the terminal side (100).

2. The system for generating personalized intervention plans based on a comprehensive assessment of adolescent development according to claim 1, wherein the terminal side (100) includes: Physiological behavior data acquisition module (110) for collecting physiological behavior data of adolescents, including EEG data and HRV data; The core cognitive ability data collection module (120) for collecting core cognitive ability data of adolescents includes visual perception test data, visual perception test data, auditory perception test data, auditory perception test data, selective attention data, sustained attention data, alternating attention data, distributed attention and reactive attention test data, sensory memory test data, working memory test data and long-term memory test data. A physical fitness data collection module (130) for collecting physical fitness data of adolescents, including daily physical fitness test scores in schools; A psychological data collection module (140) for collecting psychological data on adolescents, including mental health and personality traits; A transceiver module for sending data to the cloud platform (200) and for the cloud platform (200) to successfully receive the personalized intervention plan.

3. The system for generating personalized intervention programs based on comprehensive assessment of adolescent development according to claim 1, wherein the multidimensional assessment model (220) is a multi-level progressive mapping model established by nonlinear functions g and f, ranging from physiological aspects including brain cognition and physical fitness to psychological aspects including psychological resilience and mental health, and further to practical abilities, social abilities and multiple potentials.

4. The system for generating personalized intervention plans based on comprehensive assessment of adolescent development according to claim 3, wherein the multi-level progressive mapping model includes: The physiological basal layer used to assess core cognitive abilities and physical fitness; A psychological regulation layer is obtained by mapping the physiological basis layer through a nonlinear function g. The psychological regulation layer is used to analyze quantitative indicators of psychological resilience, quantitative indicators of mental health status, and quantitative indicators of personality traits. A capability performance layer is obtained by mapping the psychological adjustment layer through a nonlinear function f. This capability performance layer is used to predict practical ability, social ability, multiple potentials, and personality traits. A multi-dimensional potential layer is obtained by mapping the psychological adjustment layer through a nonlinear function f. This multi-dimensional potential layer is used to predict the potential of linguistic intelligence, logical-mathematical intelligence, spatial intelligence, bodily-kinesthetic intelligence, musical intelligence, interpersonal intelligence, intrapersonal cognitive intelligence, and natural cognitive intelligence.

5. The system for generating personalized intervention programs based on a comprehensive assessment of adolescent development as described in claim 4, wherein the core cognitive abilities assessed in the physiological basal layer include vision, hearing, attention, and memory; and the physical qualities assessed in the physiological basal layer include sleep quality, cardiopulmonary function, motor coordination, height, weight, NMI index, and five basic items.

6. The system for generating personalized intervention programs based on comprehensive assessment of adolescent development according to claim 4 uses feature alignment algorithm and PCA dimensionality reduction to perform multimodal data fusion on the heterogeneous data of brain core cognitive ability, HRV, physical fitness, psychological resilience, mental health and multiple potential output by the multidimensional assessment model (220) to obtain multimodal fused data.

7. The system for generating personalized intervention programs based on comprehensive assessment of adolescent development according to claim 6, wherein the prediction module (230) processes the multimodal fusion data through logistic regression and K-means clustering to predict mental health risks and potential development trajectories.

8. The system for generating personalized intervention programs based on comprehensive assessment of adolescent development according to claim 7, wherein the intervention strategy generation module (240) generates personalized course recommendations and psychological adjustment programs based on reinforcement learning (RL).

9. A method for establishing a comprehensive assessment model for adolescent development, comprising: A multi-level progressive mapping model is established using nonlinear functions g and f, ranging from physiological aspects including brain cognition and physical fitness, to psychological aspects including psychological resilience and mental health, and further to practical abilities, social abilities, and multiple potentials.

10. The method for establishing a comprehensive assessment model for adolescent development according to claim 9, wherein the multi-level progressive mapping model comprises: The physiological basal layer used to assess core cognitive abilities and physical fitness; A psychological regulation layer is obtained by mapping the physiological basis layer through a nonlinear function g. The psychological regulation layer is used to analyze quantitative indicators of psychological resilience, quantitative indicators of mental health status, and quantitative indicators of personality traits. A capability performance layer is obtained by mapping the psychological adjustment layer through a nonlinear function f. This capability performance layer is used to predict practical ability, social ability, multiple potentials, and personality traits. A multi-dimensional potential layer is obtained by mapping the psychological adjustment layer through a nonlinear function f. This multi-dimensional potential layer is used to predict the potential of linguistic intelligence, logical-mathematical intelligence, spatial intelligence, bodily-kinesthetic intelligence, musical intelligence, interpersonal intelligence, intrapersonal cognitive intelligence, and natural cognitive intelligence.

Images