AI large model-based student psychological interview auxiliary method and system

Abstract

The invention relates to the technical field of psychological analysis, in particular to a student psychological interview assisting method and system based on an AI large model, and the method comprises the following steps: obtaining a student answer text, carrying out emotion vocabulary marking by using the AI large model, recognizing a grammatical role and an emotion direction of emotion vocabularies in the text, and carrying out the quantitative analysis of the intensity of each emotion word. And evaluating the emotion direction and intensity of each sentence of the student to obtain emotion state information. According to the method, the AI large model is introduced to perform emotion analysis on the answer text of the student, so that emotion evaluation is more comprehensive and detailed, the rate and amplitude of emotion fluctuation of the student are monitored in real time, the emotion fluctuation curve is drawn, the duration time is recorded, dynamic data support is provided for psychological health evaluation, the accuracy and real-time performance of psychological evaluation are improved, and the psychological evaluation efficiency is improved. The psychological health risk of the student is accurately identified by combining emotion fluctuation and psychological pressure analysis, so that the evaluation result is more personalized, and targeted and prospective support is provided for psychological experts.

Description

Technical Field

[0001] The present invention relates to the field of psychological analysis technology, and in particular to a student psychological interview auxiliary method and system based on an AI big model. Background Art

[0002] The field of psychological analysis technology includes interdisciplinary research in psychology and computer science, and aims to evaluate, understand and intervene in the psychological state of individuals or groups through scientific analysis methods, involving multiple aspects such as mental health assessment, emotion recognition, psychological counseling and intervention. Combined with the application of natural language processing and deep learning technology, it helps analyze the psychological condition of individuals, and assists psychology experts in providing more efficient psychological services on a larger scale through the analysis of multidimensional data such as language, expression, and behavior. Psychological analysis technology is not limited to traditional questionnaires and interviews, but also gradually incorporates intelligent means, such as AI models and emotional computing, to promote the intelligent and personalized development of mental health intervention.

[0003] Among them, the student psychological interview assistance method based on AI big model refers to a method of assisting the analysis of students' psychological conditions by using large-scale artificial intelligence models. It mainly involves assisting students' psychological interviews through AI big models to improve the efficiency and accuracy of information processing during the interview. Specifically, this method analyzes students' language and emotional expressions in the interview, uses natural language processing technology to extract relevant information about their psychological state from students' speech, and generates targeted questions or feedback. The implementation of this method includes semantic understanding, sentiment analysis, and personalized question generation of conversations during the interview, so as to provide more support for psychological experts and improve the assessment of students' mental health.

[0004] Traditional student psychological interview auxiliary methods have limitations in data analysis, insufficient dynamic analysis of emotion recognition and emotional fluctuations, and rely on questionnaires and interviews. They can only provide static data on students' emotions and lack in-depth analysis of the rate of emotional change and the amplitude of fluctuations. It is difficult to accurately reflect the actual situation of students' emotional fluctuations during the interview. Emotional analysis is limited to a single dimension, and students' emotions are classified and evaluated through simple emotional vocabulary, ignoring the complexity and diversity of students' emotional expressions, resulting in a decrease in the accuracy of evaluation results. The assessment of psychological stress relies on students' self-reports and fails to combine actual emotional fluctuation data, making it difficult to comprehensively assess students' mental health status. When dealing with students' mental health issues, they face problems such as insufficient data analysis and inaccurate evaluation, and are unable to detect potential psychological risks in a timely manner. Summary of the invention

[0005] In order to solve the technical problems existing in the prior art, the embodiment of the present invention provides a student psychological interview auxiliary method and system based on an AI big model. The technical solution is as follows: In order to achieve the above purpose, the present invention adopts the following technical scheme, which is a student psychological interview auxiliary method based on AI big model, including the following steps: S1: Obtain the student's answer text, use the AI ​​big model to annotate emotional words, identify the grammatical role and emotional direction of the emotional words in the text, conduct quantitative analysis on the intensity of each emotional word, evaluate the emotional direction and intensity of each sentence of the student, and obtain emotional state information; S2: Based on the emotional state information, calculate the rate and amplitude of the student's emotional fluctuation in each round of dialogue, draw a curve chart of the student's emotional fluctuation, monitor the changing trend of the student's emotions, record the duration and amplitude of the emotional fluctuation, and generate emotional fluctuation trend data; S3: Based on the emotion fluctuation trend data, according to the change rate and amplitude of the student's emotion fluctuation, the question matching parameters are adjusted to obtain a question parameter adjustment record; S4: combining the question parameter adjustment record, extracting keywords related to life pressure, study pressure, and social pressure from students' answers, calculating emotional load values ​​of various pressure types, evaluating students' psychological pressure, and obtaining a pressure evaluation value; S5: Based on the stress assessment value, according to the psychological stress and mood fluctuation data, calculate the student's mental health risk value, mark the students with mental health risks, and obtain the psychological risk assessment result.

[0006] As a further scheme of the present invention, the emotional state information includes the grammatical role of emotional vocabulary, the emotional direction of emotional vocabulary, and the intensity of emotional vocabulary; the emotional fluctuation trend data includes the rate of emotional fluctuation, the amplitude of emotional fluctuation, and the duration of emotional fluctuation; the question parameter adjustment record includes the question tone adjustment value, the question depth adjustment value, and the question complexity adjustment value; the stress assessment value includes the life stress load value, the learning stress load value, and the social stress load value; the psychological risk assessment result includes the student's mental health risk value, mental health risk marker, and mental health risk category.

[0007] As a further solution of the present invention, the steps of obtaining the student's answer text, using the AI ​​big model to annotate the emotional vocabulary, identifying the grammatical role and emotional direction of the emotional vocabulary in the text, quantitatively analyzing the intensity of each emotional word, evaluating the emotional direction and intensity of each sentence of the student, and obtaining the emotional state information are as follows: S101: Obtain the student's answer text, use the AI ​​big model to annotate the text with sentiment words, identify sentiment words, calibrate the sentiment direction of sentiment words, identify the grammatical role of each sentiment word in the sentence, and generate sentiment word recognition results; S102: Based on the emotion word recognition result, the emotion direction is classified according to the context of the emotion word, and the intensity of each emotion word is quantified through the word frequency and emotion change rate of the emotion word to obtain emotion intensity quantification data; S103: Based on the emotion intensity quantification data, the emotion state of the sentence is evaluated according to the emotion direction and intensity of each emotion word, including the emotion direction and intensity value, to obtain emotion state information.

[0008] As a further solution of the present invention, based on the emotional state information, the rate and amplitude of the student's emotional fluctuation in each round of dialogue are calculated, a curve chart of the student's emotional fluctuation is drawn, the changing trend of the student's emotions is monitored, and the duration and amplitude of the emotional fluctuation are recorded. The steps of generating emotional fluctuation trend data are specifically as follows: S201: Based on the emotional state information, by detecting the emotional changes in each round of dialogue, comparing the emotional intensity values ​​of adjacent rounds, evaluating the rate of change of emotions, identifying the amplitude of emotional fluctuations, and obtaining the rate and amplitude data of emotional fluctuations; S202: based on the rate and amplitude data of the emotion fluctuation, drawing a curve graph of the student's emotion fluctuation to obtain an emotion fluctuation graph; S203: According to the emotion fluctuation graph, predict the changing trend of the curve, calculate the duration and fluctuation amplitude of each emotion fluctuation, and generate emotion fluctuation trend data.

[0009] As a further solution of the present invention, based on the emotion fluctuation trend data, according to the change rate and amplitude of the student's emotion fluctuation, the question matching parameters are adjusted, and the steps of obtaining the question parameter adjustment record are specifically as follows: S301: Based on the emotion fluctuation trend data, the degree of fluctuation of the student's emotion is evaluated by the amplitude of the change of the student's emotion fluctuation, and the complexity and length parameters of the question matching parameters are adjusted to obtain the complexity and length parameters; S302: According to the complexity and length parameters and the rate of change of the emotional fluctuation, the content and tone parameters of the question matching parameters are adjusted, including the topic, information content, and tone intensity of the question, to obtain the tone and content parameters of the question; S303: Using the question tone and content parameters, according to the changing direction of the emotion fluctuation, the emotional tendency parameters of the question matching parameters are adjusted to obtain a question parameter adjustment record.

[0010] As a further solution of the present invention, in combination with the question parameter adjustment record, by extracting keywords related to life pressure, study pressure, and social pressure from students' answers, calculating the emotional load values ​​of various pressure types, and evaluating students' psychological pressure, the steps of obtaining the pressure evaluation value are specifically as follows: S401: adjusting the record according to the question parameters, analyzing the student's answer text, extracting keywords related to life pressure, study pressure, and social pressure from the student's answer, and obtaining keyword extraction results; S402: Based on the keyword extraction result, according to the emotion intensity and frequency of multiple keywords, the emotion load of multiple stress types is calculated to generate emotion load evaluation data; S403: Calculate the student's psychological stress value based on the emotional load of various stress types in combination with the emotional load assessment data to obtain a stress assessment value.

[0011] As a further solution of the present invention, the specific formula for calculating the emotional load of multiple stress types is: ; Calculate the emotional load of stress types, and calculate the emotional load of each stress type separately; in, represents the emotional load of a certain stress type, Represents the pressure type The frequency of keywords, Representative The emotional intensity of the keywords, represents the base pressure load, Represents the number of keywords under this pressure type, represents the adjustment coefficient, Represents the pressure type An index of keywords.

[0012] As a further solution of the present invention, based on the stress assessment value, according to the psychological stress and mood fluctuation data, the mental health risk value of the student is calculated, and the mental health risk students are marked. The steps of obtaining the psychological risk assessment result are specifically as follows: S501: Based on the stress assessment value, according to the psychological stress and mood fluctuation data, calculating the mental health risk values ​​of multiple students to obtain risk score records; S502: Based on the risk score record, classify the student into a plurality of mental health risk levels according to the risk score to obtain a mental health risk level; S503: According to the mental health risk level, identify students with mental health risks, and generate psychological risk assessment results based on the students' emotional volatility and emotional loads of various stress types.

[0013] As a further solution of the present invention, the specific formula for calculating the mental health risk values ​​of multiple students is: ; Calculate risk value; in, Represents the student's mental health risk value, Representative The weight of each stress type, Representative Pressure evaluation value for each pressure type, Representative The weight of the emotional fluctuation data, Representative The intensity of emotional fluctuations of emotional fluctuation data, Represents the number of pressure types, Represents the number of emotional fluctuation data, Indicates Pressure type, Indicates Mood swing data.

[0014] On the other hand, a student psychological interview auxiliary system based on an AI big model is provided, and the system is applied to a student psychological interview auxiliary method based on an AI big model, and the system includes: The emotion annotation module obtains the text of students' answers, uses the AI ​​big model to annotate the text with emotion words, identifies the grammatical role and emotion direction of the emotion words in the text, calculates the intensity of each emotion word and evaluates the emotion direction and intensity of each sentence to obtain the emotion state information; The emotion fluctuation analysis module calculates the emotion fluctuation rate and amplitude of each round of students' dialogue based on the emotion state information, draws the emotion fluctuation curve, monitors the trend of students' emotion changes, records the duration and amplitude of emotion fluctuations, and generates emotion fluctuation trend data; The question adjustment module adjusts the question matching parameters based on the emotion fluctuation trend data and the change rate and amplitude of the student's emotion fluctuations, and generates a question parameter adjustment record; The stress assessment module combines the question parameter adjustment record, extracts keywords related to life stress, study stress, and social stress from the student's answer, analyzes the student's psychological stress through the student's stress data in each category, and generates a stress assessment value; The risk assessment module assesses students' mental health risks based on the stress assessment value and combines the emotional fluctuation data, marks students with mental health risks, and generates psychological risk assessment results.

[0015] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least: By introducing a large AI model to perform sentiment analysis on students' answer texts, sentiment assessment becomes more comprehensive and detailed, the rate and amplitude of students' mood fluctuations are monitored in real time, the mood fluctuation curve is drawn and the duration is recorded, dynamic data support is provided for mental health assessment, and the accuracy and real-time performance of psychological assessment are improved. By combining mood fluctuation with psychological stress analysis, students' mental health risks can be accurately identified, making assessment results more personalized and providing targeted and forward-looking support for psychological experts. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0017] Figure 1 It is a schematic diagram of the workflow of the present invention; Figure 2 This is a detailed flow chart of S1 of the present invention; Figure 3 This is a detailed flow chart of S2 of the present invention; Figure 4 This is a detailed flow chart of S3 of the present invention; Figure 5 This is a detailed flow chart of S4 of the present invention; Figure 6 This is a detailed flow chart of S5 of the present invention; Figure 7 It is a system flow chart of the present invention. DETAILED DESCRIPTION

[0018] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0019] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "example" in the present invention should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of the word "example" is intended to present the concept in a specific way. In addition, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or it can be either of the two.

[0020] In the embodiments of the present invention, "image" and "picture" can sometimes be used interchangeably. It should be noted that when the difference between them is not emphasized, the meanings they intend to express are the same. "of", "corresponding, relevant" and "corresponding" can sometimes be used interchangeably. It should be noted that when the difference between them is not emphasized, the meanings they intend to express are the same.

[0021] In the embodiments of the present invention, sometimes a subscript such as W1 may be written as a non-subscript such as W1. When the difference is not emphasized, the meanings to be expressed are the same.

[0022] In order to make the technical problems, technical solutions and advantages to be solved by the present invention more clear, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0023] See also Figure 1 The present invention provides a technical solution, a student psychological interview auxiliary method based on an AI large model, comprising the following steps: S1: Obtain the student's answer text, use the AI ​​big model to annotate emotional words, identify the grammatical role and emotional direction of the emotional words in the text, conduct quantitative analysis on the intensity of each emotional word, evaluate the emotional direction and intensity of each sentence of the student, and obtain emotional state information; S2: Based on the emotional state information, calculate the rate and amplitude of the student's emotional fluctuation in each round of dialogue, draw a curve chart of the student's emotional fluctuation, monitor the changing trend of the student's emotions, record the duration and amplitude of the emotional fluctuation, and generate emotional fluctuation trend data; S3: Based on the emotion fluctuation trend data, according to the change rate and amplitude of the student's emotion fluctuation, the question matching parameters are adjusted to obtain a question parameter adjustment record; S4: combining the question parameter adjustment record, extracting keywords related to life pressure, study pressure, and social pressure from students' answers, calculating emotional load values ​​of various pressure types, evaluating students' psychological pressure, and obtaining a pressure evaluation value; S5: Based on the stress assessment value, according to the psychological stress and mood fluctuation data, calculate the student's mental health risk value, mark the students with mental health risks, and obtain the psychological risk assessment result.

[0024] The emotional state information includes the grammatical role of emotional vocabulary, the emotional direction of emotional vocabulary, and the intensity of emotional vocabulary; the emotional fluctuation trend data includes the rate of emotional fluctuation, the amplitude of emotional fluctuation, and the duration of emotional fluctuation; the question parameter adjustment record includes the question tone adjustment value, the question depth adjustment value, and the question complexity adjustment value; the stress assessment value includes the life stress load value, the learning stress load value, and the social stress load value; the psychological risk assessment result includes the student's mental health risk value, mental health risk marker, and mental health risk category.

[0025] See also Figure 2 , obtain the student's answer text, use the AI ​​big model to annotate emotional words, identify the grammatical role and emotional direction of emotional words in the text, quantitatively analyze the intensity of each emotional word, evaluate the emotional direction and intensity of each sentence of the student, and obtain the emotional state information in the following steps: S101: Obtain the student's answer text, use the AI ​​big model to annotate the text with sentiment words, identify sentiment words, calibrate the sentiment direction of sentiment words, identify the grammatical role of each sentiment word in the sentence, and generate sentiment word recognition results; During the recognition of emotional vocabulary, the AI ​​model will extract emotional vocabulary from the text and identify all emotion-related vocabulary in the text. Emotional vocabulary includes adjectives, verbs, nouns, etc. These words can represent emotional tendencies in the text, such as "happy", "angry", "excited" and "frustrated". The model first matches the text vocabulary according to the dictionary or semantic vector model, and annotates the emotional vocabulary in combination with the context to further determine the emotional direction of each emotional vocabulary. For example, the "happy" in the sentence "I am very happy today" belongs to emotional vocabulary, with a positive emotional direction and a grammatical role as a predicate, describing the emotional state of the subject "I". In the calculation of the intensity of emotional vocabulary, if "happy" is modified by "very", the intensity will be enhanced, so the model quantifies the emotional intensity of emotional vocabulary. Assuming that the emotional intensity of the emotional vocabulary "happy" is 0.5, after being modified by "very", the emotional intensity increases to 0.8, which is quantified as 0.8. The formula is used: ; in, Represents emotional intensity. Represents the original emotion intensity (e.g. the initial intensity of "happy" is 0.5), Represents the strength of the modification (e.g., a "very" modification factor of 1.6).

[0026] If the initial intensity of "happy" is 0.5 and the intensity of the modifier "very" is 1.6, the sentiment intensity is calculated as: .

[0027] S102: Based on the emotion word recognition result, the emotion direction is classified according to the context of the emotion word, and the intensity of each emotion word is quantified through the word frequency and emotion change rate of the emotion word to obtain emotion intensity quantification data; In the process of classifying the emotional direction of sentimental vocabulary, the AI ​​model will classify the emotional direction of sentimental vocabulary according to the context. For example, in the sentence "I am very happy today", "happy" expresses positive emotions and the emotional direction is positive; if the sentence is changed to "I am not very happy today", the emotional direction will turn negative. In this process, the model uses the word frequency analysis of sentimental vocabulary and combines the context to judge the rate of change of sentimental vocabulary. For example, if "happy" appears frequently and reinforcing words such as "super" appear in subsequent sentences, the intensity of the emotion increases and the emotional direction remains positive. Assuming that the frequency of the emotional word "happy" is 3 times and the emotional intensity is 0.6 each time, the word frequency of the sentimental vocabulary can be used to judge the change in the intensity of the emotion. Use the formula: ; in, represents the average intensity of sentiment words, Represents the number of occurrences of sentiment words, Representative The intensity of the emotion words that appear.

[0028] Assume that "happy" appears 3 times, with intensities of 0.6, 0.7, and 0.8 respectively. Calculate its average intensity: F=1 / 3*(0.6+0.7+0.8)=1 / 3*2.1=0.7 S103: Based on the emotion intensity quantification data, the emotion state of the sentence is evaluated according to the emotion direction and intensity of each emotion word, including the emotion direction and intensity value, to obtain emotion state information; In the process of evaluating the emotional state, the AI ​​model combines the emotional direction and intensity of each emotional word to evaluate the emotional state of the entire sentence. If the sentence contains multiple emotional words, the model will calculate the emotional direction and intensity of the entire sentence by weighted averaging. For example, there are two emotional words "happy" and "angry" in the sentence. Assume that the emotional intensity of "happy" is 0.8, the emotional direction is positive, and the emotional intensity of "angry" is 0.4, and the emotional direction is negative. The model takes a weighted average based on the intensity and direction of these words to obtain the emotional intensity of the sentence. Assuming that the weight of the emotional direction is 0.7 for positive emotions and 0.3 for negative emotions, then according to the weighted calculation, the emotional state is positive. Use the formula: ; in, represents the intensity of the emotional state of the sentence, and represents the weight of positive and negative emotions, and Represents the intensity of positive and negative emotions.

[0029] Assuming that the positive emotion weight is 0.7, the negative emotion weight is 0.3, and the intensity of "happy" is 0.8, the emotional direction is positive, and the intensity of "angry" is 0.4, the emotional direction is negative, then calculate the emotional state of the sentence: ; Therefore, the emotional state intensity of this sentence is 0.68, which indicates positive emotion.

[0030] See also Figure 3 Based on the emotional state information, the rate and amplitude of the student's emotional fluctuation in each round of dialogue are calculated, a curve chart of the student's emotional fluctuation is drawn, the changing trend of the student's emotions is monitored, and the duration and amplitude of the emotional fluctuation are recorded. The steps of generating emotional fluctuation trend data are specifically as follows: S201: Based on the emotional state information, by detecting the emotional changes in each round of dialogue, comparing the emotional intensity values ​​of adjacent rounds, evaluating the rate of change of emotions, identifying the amplitude of emotional fluctuations, and obtaining the rate and amplitude data of emotional fluctuations; In the process of detecting emotional changes, the AI ​​model will compare the emotional intensity values ​​in each round of conversation, and evaluate the rate and amplitude of emotional changes by calculating the difference in emotional intensity between adjacent rounds. First, the model will calculate the change in emotional intensity between adjacent conversation rounds. If the emotional intensity of the current round increases compared to the previous round, the rate of emotional change is positive, otherwise it is negative. During the calculation process, the model will determine the rate of change of emotions in each round based on the intensity of the emotional vocabulary. For example, if the emotional intensity of the first round is 0.5 and the second round is 0.7, the rate of change is 0.2. Next, the model will calculate the amplitude of emotional fluctuations, and the amplitude is calculated based on the maximum change in emotional intensity between adjacent rounds. If the difference in emotional intensity between the first round and the second round is 0.4, and the change is greater in subsequent rounds, the amplitude data is the maximum change. This process ensures the accurate measurement of the rate and amplitude of emotional fluctuations, and obtains accurate emotional fluctuation data. The formula is used: ; in, Represents the rate of change of emotion, Representative The emotional intensity of the conversation, Representative The emotional intensity of the conversation.

[0031] If the emotional intensity in the first round is 0.5 and the emotional intensity in the second round is 0.7, the rate of change is: .

[0032] S202: based on the rate and amplitude data of the emotion fluctuation, drawing a curve graph of the student's emotion fluctuation to obtain an emotion fluctuation graph; In the process of drawing the emotion fluctuation curve, the AI ​​model will draw an emotion fluctuation graph based on the emotion intensity in each round of conversation. First, the model collects the emotion intensity data of each round of conversation, arranges them in chronological order, and then maps these data points into a graph to form an emotion fluctuation curve. Each data point represents the emotion intensity in a round of conversation, the horizontal axis represents the conversation round, and the vertical axis represents the emotion intensity. The model reflects the changing trend of emotion intensity on the curve to help analyze the law of emotion fluctuations. For example, if the emotion intensity gradually increases in the first few rounds and gradually decreases in subsequent rounds, the fluctuation graph will show a trend of rising and then falling. Based on these emotion fluctuation curves, the system can further analyze the trend of emotion fluctuations and predict possible emotional changes in the future. Using the formula: ; in, Represents the total intensity of emotion, Represents the number of dialogue turns, Representative The intensity of emotion in the conversation.

[0033] Assuming there are 5 rounds of dialogue, the emotional intensities are 0.6, 0.7, 0.8, 0.7, and 0.6 respectively, then the total emotional intensity is: ; Based on these data, the corresponding emotion fluctuation curve can be drawn.

[0034] S203: predicting the changing trend of the curve according to the emotion fluctuation graph, and calculating the duration and fluctuation amplitude of each emotion fluctuation to generate emotion fluctuation trend data; In the process of generating emotional fluctuation trend data, the AI ​​model analyzes the changing trend of the emotional fluctuation curve and calculates the duration and amplitude of each emotional fluctuation. First, the model calculates the emotional intensity of each round of conversation and analyzes the duration of its rise or fall. If the emotional intensity remains within a certain range, the fluctuation is considered to be continuous. Next, the model calculates the amplitude of the emotional fluctuation, which is calculated by the difference between the maximum emotional intensity and the minimum emotional intensity. If the emotional intensity increases from 0.5 to 0.8 and then falls back to 0.6 within a certain period of time, the amplitude of the fluctuation is 0.3 (0.8-0.5). Based on these data, the AI ​​model can accurately generate emotional fluctuation trend data and predict future emotional fluctuations. Using the formula: ; in, represents the fluctuation range, Represents the maximum emotional intensity over a period of time. Represents the minimum intensity of emotion over a period of time.

[0035] If the emotional intensity is 0.5, 0.6, 0.8, and 0.7 respectively during a certain period of time, the amplitude of the fluctuation is: ; Therefore, the amplitude of this fluctuation is 0.3, and the system is able to generate a sentiment fluctuation trend based on this data.

[0036] See also Figure 4 Based on the emotion fluctuation trend data, the question matching parameters are adjusted according to the change rate and amplitude of the student's emotion fluctuations, and the steps of obtaining the question parameter adjustment record are specifically as follows: S301: Based on the emotion fluctuation trend data, the degree of fluctuation of the student's emotion is evaluated by the amplitude of the change of the student's emotion fluctuation, and the complexity and length parameters of the question matching parameters are adjusted to obtain the complexity and length parameters; In the process of evaluating the trend data of emotional fluctuations, the AI ​​model will evaluate the degree of student emotional fluctuations by the amplitude of changes in the student's emotional fluctuations. First, the model will calculate the amplitude of emotional fluctuations. If the emotional fluctuations are large, it means that the student's emotional fluctuations are strong. Next, the model will adjust the complexity and length parameters of the problem matching according to the degree of emotional fluctuations. The adjustment of the complexity parameter is based on the amplitude of emotional fluctuations. If the fluctuation amplitude is large, the complexity of the problem should be increased, otherwise it should be reduced; the adjustment of the length parameter is based on the duration of the fluctuation. If the fluctuation duration is long, the length of the problem should be increased accordingly. Through this process, the model realizes the dynamic adjustment of the problem parameters to adapt to the situation of student emotional fluctuations. Specifically, if the student's emotional fluctuations change from 0.3 (smaller) to 0.8 (larger), the model will increase the complexity of the problem from 5 to 7 and the length from 10 to 15. In the end, the adjusted complexity and length parameters are 7 and 15. The formula is used: ; in, Represents the complexity of the problem, represents the complexity coefficient, Represents the amplitude of emotional fluctuations, Represents the basic complexity benchmark value.

[0037] If the mood swings , complexity coefficient , the benchmark complexity , then we can calculate: ; Therefore, the complexity of the final problem is 7.

[0038] S302: According to the complexity and length parameters and the rate of change of the emotional fluctuation, the content and tone parameters of the question matching parameters are adjusted, including the topic, information content, and tone intensity of the question, to obtain the tone and content parameters of the question; According to the rate of change of emotional fluctuations, the AI ​​model further adjusts the content and tone parameters of the question matching parameters. First, the model calculates the rate of change of emotional fluctuations and determines the speed of emotional change by comparing the changes in emotional intensity in adjacent rounds. If the rate of emotional fluctuations is large, the content of the question should increase the amount of information and strengthen the intensity of the tone; if the fluctuation rate is small, the content of the question should be reduced and the tone should be milder. For example, if the rate of emotional fluctuations is greater than 0.5, the topic of the question will become a more complex question, the amount of information will increase, and the tone will be adjusted to a stronger tone; if the rate of emotional fluctuations is less than 0.2, the tone should become calm and gentle. If the rate of change of emotional fluctuations is 0.6, the model will choose a more complex topic, more information and a stronger tone, and if the rate of emotional fluctuations is 0.1, it will choose a simple question and a gentle tone. Use the formula: ; in, represents the rate of emotional fluctuation, Representative The emotional intensity of the conversation, Represents the time difference between adjacent rounds.

[0039] If the first round of emotional intensity , Round 2 Emotional Intensity , time difference between adjacent rounds , then the mood fluctuation rate is: ; Therefore, the sentiment fluctuation rate is 0.3, which means that the sentiment fluctuation rate is small, and the model will choose relatively simple topics, less information, and use a mild tone.

[0040] S303: using the question tone and content parameters, adjusting the emotional tendency parameters of the question matching parameters according to the changing direction of the emotional fluctuation, and obtaining a question parameter adjustment record; Based on the emotional fluctuation graph, the AI ​​model will further adjust the emotional tendency parameters of the question matching parameters according to the direction of the emotional fluctuation. First, the model will identify the direction of the emotional fluctuation. If the emotional fluctuation changes positively, the emotional tendency of the question will also tend to be positive; if the emotional fluctuation changes negatively, the emotional tendency of the question will tend to be negative. Next, the model will adjust the question parameters according to the direction and intensity of the emotional fluctuation. For example, if the direction of the emotional fluctuation is positive and the intensity is large, the emotional tendency parameters of the question will be set to a higher positive value, otherwise, it will be adjusted to a stronger negative value. Assuming that the direction of the emotional fluctuation is negative and the fluctuation intensity is 0.7, the emotional tendency of the question will be adjusted to a negative intensity, such as -0.8, to reflect the student's emotional changes. Use the formula: ; in, represents the sentiment tendency parameter, represents the propensity coefficient, Represents the direction of the emotion fluctuation (positive or negative).

[0041] If the emotional fluctuation direction is negative, the intensity is 0.7, and the tendency coefficient , then the sentiment tendency parameter is: ; Therefore, the sentiment tendency parameter of the question is negative intensity -0.84.

[0042] See also Figure 5 , combined with the question parameter adjustment record, by extracting keywords related to life pressure, study pressure, and social pressure from students' answers, calculating the emotional load values ​​of various pressure types, and evaluating students' psychological pressure, the steps of obtaining the pressure evaluation value are as follows: S401: adjusting the record according to the question parameters, analyzing the student's answer text, extracting keywords related to life pressure, study pressure, and social pressure from the student's answer, and obtaining keyword extraction results; During the analysis of the students' answer text, the AI ​​model will first extract keywords related to life pressure, study pressure, and social pressure. The first step in keyword extraction is to filter out words that may reflect stress sources from the students' answer text, such as "anxiety", "tension", "exam", "work pressure", "interpersonal relationship", etc. The AI ​​model identifies these keywords by performing word segmentation and sentiment analysis on the text, and further determines their association with various types of pressure based on the context. For example, if "exam" and "tension" appear in the answer, the system can classify it as study pressure; if "interpersonal relationship" and "anxiety" appear, it is classified as social pressure. Next, the model calculates the emotional intensity of each keyword to determine the student's emotional performance under different types of pressure. Assuming that the student mentioned "anxiety" 3 times in his answer with an emotional intensity of 0.7, and mentioned "exam" once with an emotional intensity of 0.8, the model will record and summarize the frequency and intensity of each keyword, and finally obtain the keyword extraction results. Use the formula: ; in, represents the emotional load of the keyword, Representative The frequency of keywords, Representative The emotional intensity of the keywords, Represents the number of keywords.

[0043] If "anxiety" appears 3 times, the emotional intensity is 0.7, and "exam" appears once, the emotional intensity is 0.8, the calculation is: ; Therefore, the affective load is 2.9.

[0044] S402: Based on the keyword extraction result, according to the emotion intensity and frequency of multiple keywords, the emotion load of multiple stress types is calculated to generate emotion load evaluation data; The specific formula for calculating the emotional load of various stress types is: ; Calculate the emotional load of stress types, and calculate the emotional load of each stress type separately; in, represents the emotional load of a certain stress type, Represents the pressure type The frequency of keywords, Representative The emotional intensity of the keywords, Represents the number of keywords under this pressure type, represents the adjustment coefficient, represents the base pressure load, Represents the pressure type An index of keywords.

[0045] formula: ; Detailed explanation of the formula and the process of formula calculation and derivation: Formulas were used to calculate the affective load of different stress types, and the results were used to assess the impact of different types of stress on students’ emotions and help identify students’ psychological states.

[0046] Parameter meaning and setting value: The emotional load representing a certain stress type reflects the cumulative effect of the emotional intensity and frequency of all relevant keywords under a specific stress type.

[0047] Represents the pressure type The frequency of the keyword indicates the first The number of times a keyword appears in the student's answer text.

[0048] Representative The sentiment intensity of a keyword reflects the intensity of the emotion conveyed by the keyword in the text.

[0049] Represents the number of keywords under this pressure type, indicating the total number of all related keywords involved in the calculation.

[0050] stands for adjustment factor, which is used to adjust the scale of emotional load to ensure that it is within a reasonable range when compared with the baseline stress load.

[0051] It stands for baseline stress load, which indicates the baseline value of emotional load under standard stress type.

[0052] Assume that for life stress, the text analysis results are as follows: , the keyword "anxiety" appeared 4 times, emotional intensity ,frequency , the keyword "tension" appeared 3 times, emotional intensity , , there are two keywords "anxiety" and "tension", the adjustment coefficient , reference pressure load .

[0053] Substitute the parameters into the formula for calculation: ; ; ; ; P life =0.3333+0.0189; P life =0.3522; Results life =0.3522 indicates that the emotional load of life stress in this calculation is 0.3522. The result reflects the weighted sum of the emotional intensity and frequency of all relevant keywords under the life stress type, and the adjusted comprehensive emotional load.

[0054] S403: Calculating the psychological stress value of the student according to the emotional load of multiple stress types in combination with the emotional load assessment data to obtain a stress assessment value; Combined with the emotional load assessment data, the AI ​​model further calculates the students' psychological stress values ​​based on the emotional loads of different stress types. First, the model will calculate the total emotional load of each type of stress, and then calculate the students' comprehensive psychological stress values ​​based on these emotional loads. If the emotional load of life stress and social stress is high, the psychological stress value will also be high. During the calculation, the model calculates the weighted sum based on the weight of each type of stress. For example, life stress accounts for 40% of the weight, social stress accounts for 30%, and study stress accounts for 30%. The model obtains the final stress assessment value by multiplying each type of emotional load with its corresponding weight and taking the weighted sum. If the life stress load is 2.1, the social stress load is 1.2, and the study stress load is 1.5, the psychological stress value is calculated as: ; in, represents the total pressure assessment value, , , Represents the weights of life pressure, social pressure and study pressure, , , Represents the emotional load of each stress type.

[0055] Assuming that the emotional loads of life stress, social stress, and study stress are 2.1, 1.2, and 1.5 respectively, and the weights are 40%, 30%, and 30% respectively, the total stress assessment value is: ; Therefore, the student's stress assessment value is 1.65.

[0056] See also Figure 6Based on the stress assessment value, according to the psychological stress and mood fluctuation data, the mental health risk value of the student is calculated, and the mental health risk students are marked. The specific steps of obtaining the psychological risk assessment result are: S501: Based on the stress assessment value, according to the psychological stress and mood fluctuation data, calculating the mental health risk values ​​of multiple students to obtain risk score records; The specific formula for calculating the mental health risk values ​​of multiple students is: ; Calculate risk value; in, Represents the student's mental health risk value, Representative The weight of each stress type, Representative Pressure evaluation value for each pressure type, Representative The weight of the emotional fluctuation data, Representative The intensity of emotional fluctuations of emotional fluctuation data, Represents the number of pressure types, Represents the number of emotional fluctuation data, Indicates Pressure type, Indicates Mood swing data.

[0057] formula: ; Detailed explanation of the formula and the process of formula calculation and derivation: The formula is used to calculate the student’s mental health risk value; Parameter meaning and setting value: Representative The weight of each stress type reflects the impact of different stress types on mental health. Represents the weight of life stress, represents the weight of learning pressure, Represents the weight of social pressure.

[0058] Representative The stress evaluation value of each student represents the stress level of each student in each stress dimension. Assume Represents the evaluation value of a student's life stress. Represents the student's assessment of learning pressure. Represents the student's assessment value on social pressure.

[0059] Representative The weight of the emotional fluctuation data reflects the impact of emotional fluctuations on students' mental health. , .

[0060] Representative The emotional fluctuation intensity of the emotional fluctuation data indicates the magnitude of the student's emotional changes. , .

[0061] Formula calculation derivation process: ; ; ; ; ; ; ; ; ; ; The calculated mental health risk value is 3.2, which indicates the student's mental health risk level under the combined influence of stress and emotional fluctuations. It shows that the student has a high risk of mental health and needs to take appropriate intervention measures to relieve stress and stabilize emotions. Through further quantitative analysis and comparison, the risk level is further divided according to the preset thresholds, which helps managers to identify students who need special attention in a timely manner.

[0062] S502: Based on the risk score record, classify the student into a plurality of mental health risk levels according to the risk score to obtain a mental health risk level; Based on the risk score records, the AI ​​model classifies students into multiple mental health risk levels according to their risk scores. First, the model divides students into different risk levels based on the calculated risk scores. For example, if the risk score is less than 0.5, the student is at a low risk level; if the risk score is between 0.5 and 1.5, it is a medium risk level; if the risk score is greater than 1.5, it is a high risk level. The model automatically classifies and assigns the corresponding mental health risk level based on the specific risk score of each student. This process divides the risk level by setting thresholds, and determines the level to which each student belongs by comparing their risk scores. Assuming a student's risk score is 1.22, according to the set threshold, the student will be classified as a medium risk level. Use the formula: ; in, Represents the student's mental health risk level, Represents the student's mental health risk value.

[0063] If the risk value is 1.22, then according to the above formula, the student’s risk level is medium risk level.

[0064] S503: Identify students with mental health risks according to the mental health risk level, and generate a mental risk assessment result based on the students' emotional volatility and emotional load of various stress types; Based on the mental health risk level, the AI ​​model identifies students with mental health risks. First, the model will identify students at high and medium risk levels as those who need special attention based on each student's mental health risk level. Then, based on the students' emotional volatility and the emotional load of various stress types, the model will generate a psychological risk assessment result for each student. During the assessment process, the model will conduct a comprehensive analysis of emotional fluctuations and the emotional load of each type of stress. If a student has greater emotional volatility and higher emotional loads of multiple types of stress, the results of his psychological risk assessment will be more serious. For example, if a student's emotional loads in life pressure, study pressure, and social pressure are 2.5, 1.8, and 1.2 respectively, and the intensity of emotional fluctuations is 0.9, the system will generate a psychological risk assessment result for the student based on these data.

[0065] See also Figure 7 , A student psychological interview assistance system based on AI big model, a student psychological interview assistance system based on AI big model is used to execute the above-mentioned student psychological interview assistance method based on AI big model, and the system includes: The emotion annotation module obtains the text of students' answers, uses the AI ​​big model to annotate the text with emotion words, identifies the grammatical role and emotion direction of the emotion words in the text, calculates the intensity of each emotion word and evaluates the emotion direction and intensity of each sentence to obtain the emotion state information; The emotion fluctuation analysis module calculates the emotion fluctuation rate and amplitude of each round of students' dialogue based on the emotion state information, draws the emotion fluctuation curve, monitors the trend of students' emotion changes, records the duration and amplitude of emotion fluctuations, and generates emotion fluctuation trend data; The question adjustment module adjusts the question matching parameters based on the emotion fluctuation trend data and the change rate and amplitude of the student's emotion fluctuations, and generates a question parameter adjustment record; The stress assessment module combines the question parameter adjustment record, extracts keywords related to life stress, study stress, and social stress from the student's answer, analyzes the student's psychological stress through the student's stress data in each category, and generates a stress assessment value; The risk assessment module assesses students' mental health risks based on the stress assessment value and combines the emotional fluctuation data, marks students with mental health risks, and generates psychological risk assessment results.

[0066] The above embodiments can be implemented in whole or in part by software, hardware (such as circuits), firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present invention is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center by wired (such as infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (for example, a floppy disk, a hard disk, a tape), an optical medium (for example, a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0067] It should be understood that the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. A and B can be singular or plural. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship, but it may also indicate an "and / or" relationship. Please refer to the context for specific understanding.

[0068] In the present invention, "at least one" means one or more, and "plurality" means two or more. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b, or c can mean: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple.

[0069] It should be understood that in various embodiments of the present invention, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0070] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0071] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described equipment, devices and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0072] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0073] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0074] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0075] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, and other media that can store program codes.

[0076] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A student psychological interview auxiliary method based on AI big model, characterized by: The method comprises: Obtain the student's answer text, use the AI ​​big model to annotate emotional words, identify the grammatical role and emotional direction of the emotional words in the text, conduct quantitative analysis on the intensity of each emotional word, evaluate the emotional direction and intensity of each sentence of the student, and obtain emotional state information; Based on the emotional state information, calculate the rate and amplitude of the student's emotional fluctuation in each round of dialogue, draw a curve chart of the student's emotional fluctuation, monitor the changing trend of the student's emotions, record the duration and amplitude of the emotional fluctuation, and generate emotional fluctuation trend data; Based on the emotion fluctuation trend data, the question matching parameters are adjusted according to the change rate and amplitude of the student's emotion fluctuation to obtain a question parameter adjustment record; Combined with the question parameter adjustment record, keywords related to life pressure, study pressure, and social pressure are extracted from the students' answers to calculate the emotional load values ​​of various pressure types, evaluate the students' psychological pressure, and obtain a pressure evaluation value; Based on the stress assessment value, according to the psychological stress and mood fluctuation data, the student's mental health risk value is calculated, the students with mental health risks are marked, and the psychological risk assessment results are obtained.

2. The student psychological interview auxiliary method based on AI big model according to claim 1 is characterized in that: The emotional state information includes the grammatical role of emotional vocabulary, the emotional direction of emotional vocabulary, and the intensity of emotional vocabulary; the emotional fluctuation trend data includes the rate of emotional fluctuation, the amplitude of emotional fluctuation, and the duration of emotional fluctuation; the question parameter adjustment record includes the question tone adjustment value, the question depth adjustment value, and the question complexity adjustment value; the stress assessment value includes the life stress load value, the learning stress load value, and the social stress load value; the psychological risk assessment result includes the student's mental health risk value, mental health risk marker, and mental health risk category.

3. The student psychological interview auxiliary method based on AI big model according to claim 1 is characterized in that: The steps to obtain the student's answer text, use the AI ​​big model to annotate the emotional vocabulary, identify the grammatical role and emotional direction of the emotional vocabulary in the text, quantify the intensity of each emotional word, evaluate the emotional direction and intensity of each sentence of the student, and obtain the emotional state information are as follows: Obtain the student's answer text, use the AI ​​big model to annotate the text with sentiment words, identify sentiment words, calibrate the sentiment direction of sentiment words, identify the grammatical role of each sentiment word in the sentence, and generate sentiment word recognition results; Based on the emotion word recognition result, the emotion direction is classified according to the context of the emotion words, and the intensity of each emotion word is quantified through the word frequency and emotion change rate of the emotion words to obtain the emotion intensity quantification data; Based on the emotion intensity quantification data, the emotion state of the sentence is evaluated according to the emotion direction and intensity of each emotion word, including the emotion direction and intensity value, to obtain the emotion state information.

4. The student psychological interview auxiliary method based on AI big model according to claim 1 is characterized in that: Based on the emotional state information, the rate and amplitude of the student's emotional fluctuation in each round of dialogue are calculated, a curve chart of the student's emotional fluctuation is drawn, the changing trend of the student's emotions is monitored, and the duration and amplitude of the emotional fluctuation are recorded. The steps of generating emotional fluctuation trend data are specifically as follows: Based on the emotional state information, by detecting the emotional changes in each round of dialogue, comparing the emotional intensity values ​​of adjacent rounds, evaluating the rate of change of emotions, identifying the amplitude of emotional fluctuations, and obtaining the rate and amplitude data of emotional fluctuations; Based on the rate and amplitude data of the emotion fluctuation, a curve graph of the student's emotion fluctuation is drawn to obtain an emotion fluctuation graph; According to the emotion fluctuation graph, the changing trend of the curve is predicted, and the duration and fluctuation amplitude of each emotion fluctuation are calculated to generate emotion fluctuation trend data.

5. The student psychological interview auxiliary method based on AI big model according to claim 1 is characterized in that: Based on the emotion fluctuation trend data, the question matching parameters are adjusted according to the change rate and amplitude of the student's emotion fluctuations, and the steps of obtaining the question parameter adjustment record are specifically as follows: Based on the emotion fluctuation trend data, the degree of fluctuation of the student's emotion is evaluated by the amplitude of the change of the student's emotion fluctuation, and the complexity and length parameters of the question matching parameters are adjusted to obtain the complexity and length parameters; According to the complexity and length parameters, and according to the rate of change of the emotional fluctuation, the content and tone parameters of the question matching parameters are adjusted, including the topic, information content, and tone intensity of the question, to obtain the tone and content parameters of the question; The question tone and content parameters are used to adjust the emotional tendency parameters of the question matching parameters according to the changing direction of the emotional fluctuations, and a question parameter adjustment record is obtained.

6. The student psychological interview auxiliary method based on AI big model according to claim 1 is characterized in that: Combined with the question parameter adjustment record, by extracting keywords related to life pressure, study pressure, and social pressure from the students' answers, calculating the emotional load values ​​of various pressure types, and evaluating the students' psychological pressure, the steps of obtaining the pressure evaluation value are as follows: Adjust the records according to the question parameters, analyze the students' answer texts, extract keywords related to life pressure, study pressure, and social pressure from the students' answers, and obtain keyword extraction results; Based on the keyword extraction results, the emotional loads of multiple stress types are calculated according to the emotional intensity and frequency of multiple keywords to generate emotional load evaluation data; In combination with the emotional load assessment data, the psychological stress value of the student is calculated according to the emotional loads of various stress types to obtain a stress assessment value.

7. The student psychological interview auxiliary method based on AI big model according to claim 6 is characterized in that: The specific formula for calculating the emotional load of various stress types is: ； Calculate the emotional load of stress types, and calculate the emotional load of each stress type separately; in, represents the emotional load of a certain stress type, Represents the pressure type The frequency of keywords, Representative The emotional intensity of the keywords, represents the base pressure load, Represents the number of keywords under this pressure type, represents the adjustment coefficient, Represents the pressure type An index of keywords.

8. The student psychological interview auxiliary method based on AI big model according to claim 1 is characterized in that: Based on the stress assessment value, according to the psychological stress and mood fluctuation data, the mental health risk value of the student is calculated, and the mental health risk students are marked. The specific steps of obtaining the psychological risk assessment result are: Based on the stress assessment value, according to the psychological stress and mood fluctuation data, the mental health risk values ​​of multiple students are calculated to obtain risk score records; Based on the risk score records, the students are classified into a plurality of mental health risk levels according to the risk scores to obtain a mental health risk level; Based on the mental health risk level, students with mental health risks are identified, and psychological risk assessment results are generated based on the students' emotional volatility and emotional load of various stress types.

9. The student psychological interview auxiliary method based on AI big model according to claim 8 is characterized in that: The specific formula for calculating the mental health risk values ​​of multiple students is: ； Calculate risk value; in, Represents the student's mental health risk value, Representative The weight of each stress type, Representative Pressure evaluation value for each pressure type, Representative The weight of the emotional fluctuation data, E j’ Representative The intensity of emotional fluctuations of emotional fluctuation data, Represents the number of pressure types, Represents the number of emotional fluctuation data, Indicates Pressure type, Indicates Mood swing data.

10. The student psychological interview auxiliary system based on AI big model is characterized by: According to any one of claims 1 to 9, the student psychological interview auxiliary method based on the AI ​​big model comprises: The emotion annotation module obtains the text of students' answers, uses the AI ​​big model to annotate the text with emotion words, identifies the grammatical role and emotion direction of the emotion words in the text, calculates the intensity of each emotion word and evaluates the emotion direction and intensity of each sentence to obtain the emotion state information; The emotion fluctuation analysis module calculates the emotion fluctuation rate and amplitude of each round of students' dialogue based on the emotion state information, draws the emotion fluctuation curve, monitors the trend of students' emotion changes, records the duration and amplitude of emotion fluctuations, and generates emotion fluctuation trend data; The question adjustment module adjusts the question matching parameters based on the emotion fluctuation trend data and the change rate and amplitude of the student's emotion fluctuations, and generates a question parameter adjustment record; The stress assessment module combines the question parameter adjustment record, extracts keywords related to life stress, study stress, and social stress from the student's answer, analyzes the student's psychological stress through the student's stress data in each category, and generates a stress assessment value; The risk assessment module assesses students' mental health risks based on the stress assessment value and combines the emotional fluctuation data, marks students with mental health risks, and generates psychological risk assessment results.

Images