Personalized questionnaire generation and intelligent diagnosis system for psychological screening

Abstract

The invention discloses a psychological screening-oriented personalized questionnaire generation and intelligent diagnosis system, and the system comprises a data modeling investigation module which is used for collecting behavior data of a user in the aspects of psychological state, behavior characteristics and living environment; the questionnaire answering module is used for acquiring answers of the user according to the questionnaire; the intelligent analysis module is used for performing analysis according to the obtained answers and extracting data; the decision tree classification module is used for classifying the psychological state of the user according to the acquired data and obtaining a classification result; the result evaluation and treatment scheme recommendation module is used for evaluating the psychological crisis level of the user according to the classification result; and according to the psychological crisis level of the user, generating a corresponding psychological treatment scheme. The data modeling investigation module and the subsequent process show remarkable advantages in the aspects of mental health assessment and treatment, and powerful support and guarantee can be provided for mental health service.

Description

Technical Field

[0001] The present invention relates to the technical field of psychological surveys, and particularly to a personalized questionnaire generation and intelligent diagnosis system for psychological screening. Background Art

[0002] Mental health refers to a good or normal state in all aspects and activities of the mind. Mental health is an important and inseparable part of modern people's health. However, with the development of society and the accelerating pace of life, mental health problems have become increasingly prominent.

[0003] Currently, in questionnaire surveys, data usually needs to be manually collected, sorted, and analyzed. This process is relatively cumbersome and time-consuming. The results of questionnaire surveys are to a certain extent affected by the subjective factors of the investigators or respondents. For example, the way of questioning by the investigator, the understanding ability and willingness to answer of the respondent, etc. may all affect the accuracy of the results. In questionnaire surveys, if the questionnaire design is unreasonable or the respondent does not fill it out carefully, it may lead to low-quality data collection. Questionnaire surveys usually need to be conducted at specific times and places, and a large amount of manpower and material resources are required to distribute and collect the questionnaires. This limits the scope and flexibility of the survey. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a personalized questionnaire generation and intelligent diagnosis system for psychological screening, which can improve the fineness of data preprocessing to ensure the accuracy and comprehensiveness of deformation monitoring.

[0005] To solve the above technical problems, the technical solution of the present invention is as follows:

[0006] In a first aspect, a personalized questionnaire generation system for psychological screening includes:

[0007] A data modeling and survey module, configured to collect behavioral data of a user's mental state, behavioral characteristics, and living environment; preprocess the behavioral data to obtain preprocessed historical data; construct a user mental state model based on the preprocessed historical data; simulate the user's mental state as different elements in a finite element model; analyze the user's mental state according to the finite element model to obtain the user's mental state information; identify risk points of the user based on the user's mental state information; and generate a questionnaire according to the risk points;

[0008] A questionnaire answering module, configured to obtain the user's answers according to the questionnaire;

[0009] An intelligent analysis module, configured to analyze and extract data according to the obtained answers;

[0010] A decision tree classification module, which is used to classify the user's mental state according to the acquired data and obtain the classification result;

[0011] A result evaluation and treatment plan recommendation module, which is used to evaluate the user's mental crisis level according to the classification result; and generate a corresponding psychological treatment plan according to the user's mental crisis level.

[0012] Further, the mental state is the user's emotional state; the emotional state includes happiness, sadness, anger, anxiety, stress and excitement; collect the user's mental state, behavioral characteristics, and behavioral data in the living environment, including:

[0013] Collect the user's behavioral characteristics, including login frequency, browsing content, and interaction behaviors. By analyzing the user's behavioral data, identify the behavior patterns, and the behavior patterns include active at night, high-frequency interaction and preferences;

[0014] Obtain the user's living environment information according to the communication, and the environmental information includes work status, family relationship and social activities.

[0015] Further, construct a user mental state model according to the preprocessed historical data, including:

[0016] Collect the historical data. The collected data comes from chat records, emails and social media conversations; clean the collected data to remove irrelevant information, and the irrelevant information includes advertisements and non-text content;

[0017] Convert the cleaned data into a unified format text, and segment the text into words or phrases;

[0018] Convert the text into a word frequency vector, and add the inverse document frequency to the word frequency to convert the word into a vector with a fixed dimension;

[0019] Select a recurrent neural network structure according to the text, and define the loss function and optimizer for the text by using the recurrent neural network structure;

[0020] Divide the preprocessed data into a training set, a validation set and a test set, use the training set to train the neural network, and adjust the neural network model parameters through the backpropagation algorithm. Specifically, it includes: setting the initial parameters for each layer of the neural network, inputting the training data, performing forward propagation through the neural network, and calculating the output of each layer; starting from the output layer, calculate the gradient of the loss function with respect to the output of this layer; use the chain rule to calculate the gradient of the loss function with respect to the parameters of each layer layer by layer; update the parameters of the neural network by using the calculated gradient and combining the gradient descent algorithm;

[0021] Evaluate the performance of the neural network using the validation set to obtain the evaluation results, and adjust the structure or parameters of the neural network according to the evaluation results to build a neural network model;

[0022] Train the neural network model using historical data to enable it to learn and recognize the relationship between the user's mental state and communication records; evaluate the performance of the neural network model through cross-validation and adjust the hyperparameters of the neural network model to obtain the trained neural network model.

[0023] Furthermore, the calculation formula of the loss function is:

[0024] L = -[y·log(σ(z))+(1 - y)·log(1 - σ(z))];

[0025] Among them, L represents the value of the loss function; y represents the true label of the sample; z represents the output value of the model; σ(z) represents the sigmoid function.

[0026] Furthermore, analyze the user's mental state based on the finite element model, including:

[0027] Process the collected data to obtain accurate data;

[0028] Divide the data into units;

[0029] Extract features for each unit, and the features include emotional words, emojis, language style, and topic conversion frequency;

[0030] Segment the text into independent words or phrases and remove stop words; based on the results of all word segmentations collected and remove duplicate words to build a vocabulary containing all non-repeating words; convert the text into vector form according to the vocabulary, and each dimension of the vector corresponds to a word in the vocabulary; for each text, if a certain word appears in the text, the value on the corresponding dimension is the number of times the word appears in the text or is marked as 1 to indicate its existence; if it does not appear, the value is 0; each text will be converted into a vector with the same length as the vocabulary;

[0031] Based on the extracted features, establish a mental state model for each unit;

[0032] Analyze each mental state model to identify the periods of high and low user emotions to obtain the user's mental state information;

[0033] Judge whether the user has emotional out-of-control according to the user's language, tone, and reaction, and emotional out-of-control includes anger, frustration, and anxiety;

[0034] According to the user's negative emotions, use a gentle tone and words to comfort the user and express understanding and sympathy;

[0035] Based on the user's degree of emotional out - of - control and potential risks, decide whether to seek external assistance; the external assistance includes the customer service supervisor, psychologist, and security team;

[0036] Quickly contact the external assistance through phone, instant messaging tools or internal systems, and clearly explain the situation of the user's emotional out - of - control, including the user's performance, possible reasons, and current needs;

[0037] Before the external assistance arrives or intervenes, maintain communication with the user to stabilize the user's emotions.

[0038] Furthermore, based on the user's psychological state information, identify the user's risk points; generate a questionnaire according to the risk points, including:

[0039] Integrate the obtained psychological state information, and the integration of data includes chat records, social media interactions, and user feedback;

[0040] By analyzing the integrated data, obtain the risk factors existing in the user's psychological state. The risk factors include persistent low mood, anxiety tendency, and negative evaluations of the product;

[0041] Set a series of indicators according to the identified risk factors. The indicators include the duration, frequency, and intensity of negative emotions;

[0042] Compare the user's psychological state information with the indicators to classify the risk level;

[0043] Design a corresponding questionnaire according to the identified risk points.

[0044] Furthermore, analyze and extract data according to the obtained answers, including:

[0045] Organize and summarize the answers obtained from the questionnaire, and clean the data according to the summarized answers;

[0046] Based on the cleaned data, obtain answers that remove duplicates, invalidity, and obvious errors;

[0047] Conduct quantitative analysis on the cleaned data. Quantitative analysis includes multiple - choice questions and open - ended questions;

[0048] Conduct cross - analysis on the data after quantitative analysis. Cross - analysis includes the user's age, gender, and usage habits;

[0049] Identify the user's main viewpoints, needs, and pain points through cross - analysis, and judge the user's emotional tendency through sentiment analysis technology;

[0050] Based on the analyzed data, obtain key information.

[0051] Second aspect, a personalized questionnaire intelligent diagnosis method for psychological screening, including:

[0052] Collect behavioral data of the user's mental state, behavioral characteristics, and living environment; preprocess the behavioral data to obtain preprocessed historical data; construct a user mental state model based on the preprocessed historical data; simulate the user's mental state as different elements in a finite element model; analyze the user's mental state according to the finite element model to obtain the user's mental state information; identify the user's risk points based on the user's mental state information; generate a questionnaire according to the risk points;

[0053] Obtain the user's answers according to the questionnaire;

[0054] Analyze and extract data according to the obtained answers;

[0055] Classify the user's mental state according to the obtained data and obtain a classification result;

[0056] Evaluate the user's mental crisis level according to the classification result; generate a corresponding psychological treatment plan according to the user's mental crisis level.

[0057] Third aspect, a computing device, including:

[0058] One or more processors;

[0059] A storage device for storing one or more programs, which when executed by the one or more processors cause the one or more processors to implement the method.

[0060] Fourth aspect, a computer-readable storage medium storing a program that implements the method when executed by a processor.

[0061] The above solution of the present invention has at least the following beneficial effects.

[0062] In the present invention, this module can comprehensively collect behavioral data of users in multiple aspects such as mental state, behavioral characteristics, and living environment, ensuring the comprehensiveness and multi-dimensions of the data and providing a rich information basis for subsequent analysis; through data preprocessing steps, the original data can be effectively cleaned, transformed, and prepared, improving the quality and applicability of the data. This helps to eliminate noise, errors, and inconsistencies, making the data more interpretable and providing a reliable basis for subsequent modeling and analysis; based on the preprocessed historical data, a personalized user mental state model can be constructed. Such a model is closer to the actual situation of users and helps to more accurately understand and analyze the mental state of users; simulating the mental state of users as different elements in a finite element model is an innovative application method. The finite element model can more meticulously analyze the mental state of users and reveal its internal structure and dynamic changes.

[0063] In the present invention, by analyzing the finite element model, the risk points of users can be accurately identified, which helps to timely discover potential mental problems and provide strong support for subsequent intervention and treatment; the questionnaire generated according to the risk points is more targeted and efficient, can more directly touch on the key issues of users, and improves the efficiency and effectiveness of the survey; the intelligent analysis module can deeply analyze the obtained answers and extract valuable data information. Such analysis not only has depth but also can cover a wide range of aspects, providing comprehensive support for subsequent classification and evaluation. The decision tree classification module can accurately classify the mental state of users based on the obtained data. This classification method is objective and highly interpretable, helping to more precisely understand the mental state of users; the decision tree classification module can accurately classify the mental state of users based on the obtained data. This classification method is objective and highly interpretable, helping to more precisely understand the mental state of users. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 It is a schematic diagram of a personalized questionnaire generation and intelligent diagnosis system for psychological screening provided by an embodiment of the present invention.

[0065] Figure 2 It is a schematic diagram of a process for a personalized questionnaire generation and intelligent diagnosis method for psychological screening provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0066] Hereinafter, exemplary embodiments of the present disclosure will be described in more detail with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided so that the present disclosure can be more thoroughly understood and the scope of the present disclosure can be fully conveyed to those skilled in the art.

[0067] As Figure 1As shown in the figure, an embodiment of the present invention proposes a personalized questionnaire generation system for psychological screening, including:

[0068] A data modeling and investigation module, configured to collect behavioral data of a user's mental state, behavioral characteristics, and living environment; preprocess the behavioral data to obtain preprocessed historical data; construct a user mental state model according to the preprocessed historical data; simulate the user's mental state as different elements in a finite element model; analyze the user's mental state according to the finite element model to obtain the user's mental state information; identify the user's risk points based on the user's mental state information; generate a questionnaire according to the risk points;

[0069] A questionnaire answering module, configured to obtain the user's answers according to the questionnaire;

[0070] An intelligent analysis module, configured to analyze and extract data according to the obtained answers;

[0071] A decision tree classification module, configured to classify the user's mental state according to the obtained data and obtain a classification result;

[0072] A result evaluation and treatment plan recommendation module, configured to evaluate the user's mental crisis level according to the classification result; generate a corresponding psychological treatment plan according to the user's mental crisis level.

[0073] In an embodiment of the present invention, through data collection, data in multiple aspects such as the user's mental state, behavioral characteristics, and living environment are comprehensively collected. This systematic data collection method can ensure the comprehensiveness and accuracy of information, providing a solid foundation for subsequent data analysis and model construction. Through local processing of the data, the data can be cleaned, outliers and noise can be eliminated, thereby improving the data quality and the accuracy of the model. Through the construction of the model, a more refined method can be used to analyze and understand the user's mental state. Through risk assessment, the user's risk points can be identified. Through the intelligent analysis module and the decision tree classification module, the system can automatically process and analyze the user's answers and scientifically classify the user's mental state. This intelligent processing method not only improves efficiency but also increases the objectivity of classification. Through the mental state classification result, the system can evaluate its mental crisis level and generate a corresponding psychological treatment plan.

[0074] In a preferred embodiment of the present invention, the calculation formula of the loss function is:

[0075] L = -[y·log(σ(z)) + (1 - y)·log(1 - σ(z))];

[0076] Among them, L represents the value of the loss function; y represents the true label of the sample; z represents the output value of the model; and σ(z) represents the sigmoid function.

[0077] In the embodiments of the present invention, through the sigmoid function, the z output by the model is converted into a probability value, which enables the loss function to directly reflect the gap between the probability predicted by the model and the true label. This probability interpretability helps to more intuitively understand the prediction performance and optimization direction of the model. The loss function performs well in dealing with imbalanced datasets. Since it separately considers the losses of the positive class (y = 1) and the negative class (y = 0) and weights them through the logarithmic function, it can effectively train the model even in the case of imbalanced class distributions. The loss function has good properties in the gradient descent optimization process. When the model makes a wrong prediction, the gradient of the loss function is large, which helps the model quickly correct the error; while when the model's prediction is close to correct, the gradient gradually decreases, which helps the model make finer adjustments when approaching the optimal solution. It has a wide range of applications in various classification tasks. The loss function has a certain robustness to noisy data and outliers. Since it focuses on the difference in probability distributions rather than simple prediction errors, it can, to a certain extent, reduce the impact of noise and outliers on model training.

[0078] In a preferred embodiment of the present invention, analyzing the user's mental state according to a finite element model includes:

[0079] Processing the collected data to obtain accurate data;

[0080] Dividing the data into units;

[0081] Extracting features for each unit, and the features include emotional words, emojis, language style, and topic transition frequency;

[0082] Segmenting the text into independent words or phrases and removing stop words; according to the results of all word segmentations collected and removing duplicate words, constructing a vocabulary list containing all non-duplicate words; converting the text into vector form according to the vocabulary list, where each dimension of the vector corresponds to a word in the vocabulary list; for each text, if a certain word appears in the text, the value on the corresponding dimension is the number of times the word appears in the text or is marked as 1 to indicate its existence; if it does not appear, the value is 0; each text will be converted into a vector with the same length as the vocabulary list;

[0083] Based on the extracted features, establishing a mental state model for each unit;

[0084] Analyzing each mental state model to identify the high and low periods of the user's emotion, so as to obtain the user's mental state information;

[0085] Determine whether the user has lost emotional control based on the user's language, tone, and reaction. Losing emotional control includes anger, frustration, and anxiety;

[0086] According to the user's negative emotions, use a gentle tone and words to soothe the user, expressing understanding and sympathy;

[0087] Based on the degree of the user's emotional out-of-control and potential risks, select whether external assistance is needed. The external assistance includes the customer service supervisor, psychologist, and security team;

[0088] Quickly contact the external assistance through phone, instant messaging tool, or internal system, and concisely explain the situation of the user's emotional out-of-control, including the user's performance, possible reasons, and current needs;

[0089] Maintain communication with the user before the external assistance arrives or intervenes to stabilize the user's emotions.

[0090] In the embodiment of the present invention, through precise data processing, more reliable features can be extracted, thereby constructing a more accurate mental state model. Comprehensive feature extraction helps to capture the changes in the user's mental state more meticulously and improve the sensitivity of analysis. It is not only efficient but also capable of processing complex text information, providing a powerful tool for mental state analysis. The process can identify the high and low emotional periods of the user, which helps to understand the user's emotional fluctuations and change rules more deeply. When the user loses emotional control, the process can quickly respond, use a gentle tone and words to soothe the user, and express understanding and sympathy. The mechanism ensures that professional help and support can be obtained in a timely manner when necessary, improving the response ability and security of the entire system. It helps to maintain contact with the user, further stabilize the user's emotions, and ensure the smooth transmission of information.

[0091] In a preferred embodiment of the present invention, based on the user's mental state information, identify the user's risk points; generate a questionnaire according to the risk points, including:

[0092] Integrate the obtained mental state information. The integration of data includes chat records, social media interactions, and user feedback;

[0093] Analyze the integrated data to obtain the risk factors existing in the user's mental state. The risk factors include persistent low mood, anxiety tendency, and negative evaluations of the product;

[0094] Set a series of indicators according to the identified risk factors. The indicators include the duration, frequency, and intensity of negative emotions;

[0095] Compare the user's mental state information with the indicators to classify the risk level;

[0096] Design a corresponding questionnaire according to the identified risk points.

[0097] In the embodiments of the present invention, the comprehensive data collection method helps to more accurately understand the psychological state of users, providing a rich data basis for subsequent risk factor analysis and risk level classification. Accurate risk factor identification helps to timely discover problems, providing a clear direction for subsequent intervention measures. The indicators are not only specific and quantifiable, but also can comprehensively reflect the risk degree of users' psychological state, providing an objective basis for the classification of risk levels. Risk level classification helps to take corresponding intervention measures for users with different risk levels, improving the pertinence and effectiveness of services. Targeted questionnaire design can more accurately collect relevant information of users, providing strong data support for subsequent psychological state assessment and treatment plan formulation.

[0098] In a preferred embodiment of the present invention, analyze and extract data according to the obtained answers, including:

[0099] Sort out and summarize the answers obtained in the questionnaire, and clean the data according to the summarized answers;

[0100] According to the cleaned data, obtain answers with duplicates, invalidity, and obvious errors removed;

[0101] Conduct quantitative analysis on the cleaned data, and the quantitative analysis includes multiple-choice questions and open-ended questions;

[0102] Conduct cross-analysis on the data after quantitative analysis, and the cross-analysis includes the age, gender, and usage habits of users;

[0103] Identify the main opinions, needs, and pain points of users through cross-analysis, and judge the emotional tendency of users through sentiment analysis technology;

[0104] Obtain key information according to the analyzed data. By cleaning the data of the summarized answers and removing duplicate, invalid, and obviously wrong answers, the accuracy and reliability of the data are greatly improved, providing a high-quality data basis for subsequent analysis. Quantitative analysis helps to convert subjective answers into comparable numerical data, facilitating further data processing and interpretation.

[0105] In the embodiments of the present invention, at the beginning of the process, the answers in the questionnaire are sorted and summarized, which helps to integrate the scattered data into a unified data set for subsequent analysis and processing. By performing cross-analysis on the data after quantitative analysis and considering multiple dimensions such as the age, gender, and usage habits of users, it is possible to gain a deeper understanding of the characteristics and differences of the user group, providing strong support for accurate user profiling and market segmentation. By using sentiment analysis technology to judge the sentiment tendency of users, it is possible to keenly capture the emotional changes and attitudes of users, providing important references for enterprises to understand aspects such as user satisfaction, loyalty, and brand image. The entire analysis process can ultimately extract key information, which helps enterprises quickly grasp user needs and pain points, providing decision-making basis for product improvement, service optimization, and market strategy adjustment.

[0106] As Figure 2 shown, an embodiment of the present invention also provides a personalized questionnaire intelligent diagnosis method for psychological screening, including:

[0107] Constructing an investigation module for collecting behavioral data on the psychological state, behavioral characteristics, and living environment of users; preprocessing the behavioral data to obtain preprocessed historical data; constructing a user psychological state model based on the preprocessed historical data; simulating the psychological state of users as different elements in a finite element model; analyzing the psychological state of users based on the finite element model to obtain user psychological state information; identifying risk points of users based on the user psychological state information; generating a questionnaire according to the risk points;

[0108] A question answering module for obtaining the answers of users according to the questionnaire;

[0109] An analysis module for analyzing and extracting data according to the obtained answers;

[0110] A classification module for classifying the psychological state of users according to the obtained data and obtaining a classification result;

[0111] A recommendation module, a result evaluation and treatment plan recommendation module, for evaluating the psychological crisis level of users according to the classification result; generating a corresponding psychological treatment plan according to the psychological crisis level of users.

[0112] It should be noted that this system corresponds to the above method. All implementation methods in the above method embodiments are applicable to this embodiment and can achieve the same technical effects.

[0113] The above are the preferred embodiments of the present invention. It should be pointed out that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present invention.

Claims

1. A personalized questionnaire generation and intelligent diagnosis system for psychological screening, characterized in that Including: A data modeling investigation module, used to collect behavioral data on the user's mental state, behavioral characteristics, and living environment; Preprocess the behavioral data to obtain preprocessed historical data; Based on the preprocessed historical data, construct a user mental state model; Simulate the user's mental state as different elements in a finite element model; Analyze the user's mental state according to the finite element model to obtain the user's mental state information; Based on the user's mental state information, identify the user's risk points; Generate a questionnaire according to the risk points; A questionnaire answering module, used to obtain the user's answers according to the questionnaire; An intelligent analysis module, used to analyze and extract data according to the obtained answers; A decision tree classification module, used to classify the user's mental state according to the obtained data and obtain a classification result; A result evaluation and treatment plan recommendation module, used to evaluate the user's mental crisis level according to the classification result; Generate a corresponding psychological treatment plan according to the user's mental crisis level.

2. The personalized questionnaire generation and intelligent diagnosis system for psychological screening according to claim 1, characterized in that, The mental state is the user's emotional state; The emotional state includes happiness, sadness, anger, anxiety, stress, and excitement; Collect behavioral data on the user's mental state, behavioral characteristics, and living environment, including: Collect the user's behavioral characteristics, including login frequency, browsing content, and interaction behavior. By analyzing the user's behavioral data, identify behavioral patterns, and the behavioral patterns include night-time activity, high-frequency interaction, and preferences; Obtain the user's living environment information according to the communication, and the environmental information includes work status, family relationship, and social activities.

3. The personalized questionnaire generation and intelligent diagnosis system for psychological screening according to claim 2, wherein Based on the preprocessed historical data, construct a user mental state model, including: Collect historical data, and the collected data comes from chat records, emails, and social media conversations; Clean the collected data to remove irrelevant information, and the irrelevant information includes advertisements and non-text content; Convert the cleaned data into a unified format text, and cut the text into words or phrases; Convert the text into a word frequency vector, and add the inverse document frequency to the word frequency to convert the word into a vector with a fixed dimension; Select a recurrent neural network structure according to the text, and define a loss function and an optimizer for the text by using the recurrent neural network structure; Divide the preprocessed data into a training set, a validation set, and a test set. Use the training set to train the neural network, and adjust the neural network model parameters through the backpropagation algorithm. Specifically, it includes: Set initial parameters for each layer of the neural network, input the training data, perform forward propagation through the neural network, and calculate the output of each layer; Starting from the output layer, calculate the gradient of the loss function with respect to the output of this layer; Use the chain rule to calculate the gradient of the loss function with respect to the parameters of each layer layer by layer; By using the calculated gradient and combining the gradient descent algorithm, update the parameters of the neural network; Evaluate the performance of the neural network by using the validation set to obtain an evaluation result, and adjust the neural network structure or parameters according to the evaluation result to construct a neural network model; Train a neural network model using historical data to enable it to learn and recognize the relationship between the user's mental state and communication records; evaluate the performance of the neural network model through cross-validation and adjust the hyperparameters of the neural network model to obtain the trained neural network model.

4. The personalized questionnaire generation and intelligent diagnosis system for psychological screening according to claim 3, wherein The calculation formula of the loss function is: L = -[y·log(σ(z))+(1 - y)·log(1 - σ(z))]; where L represents the value of the loss function; y represents the true label of the sample; z represents the output value of the model; and σ(z) represents the sigmoid function.

5. The personalized questionnaire generation and intelligent diagnosis system for psychological screening according to claim 4, characterized in that, Analyze the user's mental state based on the finite element model, including: Process the collected data to obtain accurate data; Divide the data into elements; Extract features for each element, and the features include emotional words, emojis, language style, and topic transition frequency; Segment the text into independent words or phrases and remove stop words; based on all the segmented results collected, remove duplicate words, and construct a vocabulary list containing all non-repeating words; convert the text into vector form according to the vocabulary list, and each dimension of the vector corresponds to a word in the vocabulary list; for each text, if a word appears in the text, the value on the corresponding dimension is the number of times the word appears in the text or is marked as 1 to indicate its existence; if it does not appear, the value is 0; each text will be converted into a vector with the same length as the vocabulary list; Based on the extracted features, establish a mental state model for each element; Analyze each mental state model to identify the high and low periods of the user's emotions to obtain the user's mental state information; Judge whether the user has lost emotional control according to the user's language, tone, and reaction, and emotional out-of-control includes anger, frustration, and anxiety; According to the user's negative emotions, use a gentle tone and words to soothe the user and express understanding and sympathy; According to the degree of the user's emotional out-of-control and potential risks, select whether to seek external assistance; the external assistance sought includes the customer service supervisor, psychologist, and security team; Quickly contact the external assistance through phone, instant messaging tool, or internal system, and clearly explain the situation of the user's emotional out-of-control, including the user's performance, possible reasons, and current needs; Before the external assistance arrives or intervenes, maintain communication with the user and stabilize the user's emotions.

6. The personalized questionnaire generation and intelligent diagnosis system for psychological screening according to claim 5, wherein Based on the user's mental state information, identify the user's risk points; Generate a questionnaire according to the risk points, including: Integrate the obtained mental state information, and the integration of data includes chat records, social media interactions, and user feedback; Analyze the integrated data to obtain the risk factors existing in the user's mental state, and the risk factors include persistent low mood, anxiety tendency, and negative evaluations of the product; Set a series of indicators according to the identified risk factors, and the indicators include the duration, frequency, and intensity of negative emotions; Compare the user's mental state information with the indicators to classify the risk level; Design a corresponding questionnaire according to the identified risk points.

7. The personalized questionnaire generation and intelligent diagnosis system for psychological screening according to claim 6, characterized in that, Analyze and extract the data according to the obtained answers, including: Sort out and summarize the answers obtained from the questionnaire, and clean the data according to the summarized answers; Based on the cleaned data, obtain answers with duplicates, invalidity, and obvious errors removed; Conduct quantitative analysis on the cleaned data, which includes multiple-choice questions and open-ended questions; Conduct cross-analysis on the data after quantitative analysis, which includes the age, gender, and usage habits of users; Identify the main viewpoints, needs, and pain points of users through cross-analysis, and judge the emotional tendency of users through sentiment analysis technology; Obtain key information based on the analyzed data.

8. A personalized questionnaire generation and intelligent diagnosis method for psychological screening, characterized in that, The system is used to execute the method described in any one of claims 1 to 7, including: Collect behavioral data on the psychological state, behavioral characteristics, and living environment of users; preprocess the behavioral data to obtain preprocessed historical data; construct a user psychological state model based on the preprocessed historical data; simulate the psychological state of users as different elements in a finite element model; analyze the psychological state of users according to the finite element model to obtain user psychological state information; identify risk points of users based on the user psychological state information; generate a questionnaire according to the risk points; Obtain the answers of users according to the questionnaire; Analyze and extract data according to the obtained answers; Classify the psychological state of users according to the obtained data and obtain a classification result; Evaluate the psychological crisis level of users according to the classification result; generate a corresponding psychological treatment plan according to the psychological crisis level of users.

9. A computing device, characterized in that, Including: One or more processors; A storage device for storing one or more programs, which when executed by the one or more processors, causes the one or more processors to implement the method described in claim 8.

10. A computer-readable storage medium, characterized in that, A program is stored in the computer-readable storage medium, and when the program is executed by a processor, it implements the method described in claim 8.

Images