Personalized emotion evaluation questionnaire generation method and system based on real-time physiological data and neural network modeling

Abstract

The invention discloses a personalized emotion evaluation questionnaire generation method and system based on real-time physiological data and neural network modeling, relates to the technical field of emotion evaluation, and solves the problems of lack of personalization, data isolation and poor dynamic adaptability of an emotion evaluation questionnaire in the prior art. The method comprises the steps of obtaining multi-modal physiological data and performing dynamic feature extraction, predicting emotional state classification and emotional degree indexes of a user by utilizing an emotional state evaluation model, and automatically generating a personalized emotional evaluation questionnaire by utilizing a questionnaire generation model according to a model prediction result. And a feedback mechanism is constructed to continuously optimize the emotional state evaluation model and the questionnaire generation model, so that the problems of insufficient personalization, data isolation and poor dynamic adaptability in the prior art can be effectively overcome, automatic generation of the personalized emotional evaluation questionnaire is realized, the emotional evaluation efficiency is improved, and the user experience is improved. And the questionnaire content can be differentiated according to individual differences.

Description

Technical Field

[0001] The present application relates to the field of emotion assessment technology, and in particular to a method and system for generating a personalized emotion assessment questionnaire based on real-time physiological data and neural network modeling. Background Art

[0002] Emotional assessment has important application value in areas such as mental health, education, market research, and healthcare. Accurate and timely emotion recognition and assessment not only facilitates individual psychological intervention and emotional regulation, but also provides key support for personalized education, consumer behavior analysis, and disease-assisted diagnosis. Currently, common emotion assessment methods include:

[0003] Standardized questionnaires, such as the Positive and Negative Affect Scale (PANAS), the Beck Depression Inventory (BDI), and the State-Trait Anxiety Inventory (STAI), assess emotional states through fixed questionnaire content. These methods have the advantages of widespread application and uniform standards, but they also suffer from fixed assessment content and the inability to dynamically adjust to individual differences, resulting in a lack of personalization.

[0004] Physiological data monitoring methods collect physiological signals such as heart rate (HR), heart rate variability (HRV), galvanic skin response (GSR), and electroencephalogram (EEG) to infer an individual's emotional state. Although physiological data can provide objective indicators, relying solely on physiological data lacks effective integration with psychological characteristics and individual differences, which can easily lead to biased emotional inferences.

[0005] Manual interviewing: Professional psychologists or evaluators directly assess the emotional state of the person being assessed through interactive interviews. While this method can combine verbal and nonverbal information for comprehensive assessment, it suffers from high costs, low efficiency, and the results are significantly influenced by subjective factors.

[0006] In summary, the existing technology has the following major deficiencies:

[0007] (1) Lack of personalization: Existing standardized questionnaires use a unified set of fixed questions, which cannot be dynamically adjusted according to the physiological characteristics, psychological state and emotional fluctuation characteristics of different individuals. It is difficult to reflect the real emotional differences of individuals, resulting in insufficient individual adaptability and accuracy of the assessment results.

[0008] (2) Data isolation: In existing technologies, physiological signal data (such as heart rate, EEG, and galvanic skin response) and psychological questionnaire data are usually collected and analyzed independently. There is a lack of a unified data fusion mechanism, and a real-time correlation between physiological indicators and psychological questionnaire feedback cannot be established. As a result, it is impossible to dynamically guide questionnaire content design or adjust evaluation strategies based on physiological changes.

[0009] (3) Poor dynamic adaptability: The traditional emotion assessment process is basically static and one-way in design. It lacks real-time monitoring and feedback adjustment mechanisms. It is unable to adjust the content, order or difficulty of questions in a timely manner according to the user's emotional changes during the assessment process. This reduces the timeliness, interactivity and accuracy of emotion assessment, and is not conducive to the precise tracking of continuous and multi-point emotional states. Summary of the Invention

[0010] The purpose of this application is to overcome the problems of lack of personalization, data isolation and poor dynamic adaptability of emotion assessment questionnaires in the prior art, and to provide a method and system for generating personalized emotion assessment questionnaires based on real-time physiological data and neural network modeling.

[0011] In the first aspect, a method for generating a personalized emotion assessment questionnaire based on real-time physiological data and neural network modeling is provided, comprising:

[0012] Acquiring physiological data of the user, and preprocessing the physiological data and extracting time series dynamic features;

[0013] Inputting the temporal dynamic features into a pre-trained emotional state assessment model to obtain the user's current emotional state classification and emotional degree index;

[0014] Input the emotional state classification and emotional degree index within a specified time period into the pre-trained questionnaire generation model to obtain a personalized emotional assessment questionnaire;

[0015] The emotional state assessment model and questionnaire generation model are continuously optimized based on the behavioral feedback data generated by users when filling out personalized emotional questionnaires and the physiological data collected simultaneously.

[0016] In some possible implementations, the physiological data includes heart rate, heart rate variability, galvanic skin response, respiratory rate, pupil dilation degree, and EEG.

[0017] In some possible implementations, the preprocessing includes performing denoising, smoothing, and feature extraction on the physiological signal using wavelet transform and fast Fourier transform.

[0018] In some possible implementations, the temporal feature modeling includes inputting the preprocessed physiological data into a long short-term memory network or a Transformer model to extract temporal dynamic features.

[0019] In some possible implementations, the pre-trained emotional state assessment model includes a multimodal fusion unit, an emotion classification prediction unit, and an emotion degree quantification unit, wherein the multimodal fusion unit introduces a cross-modal attention mechanism, performs weighted fusion of different physiological signal features to learn the correlation between different physiological signals, the emotion classification prediction unit performs a comprehensive analysis of multimodal physiological data based on a deep neural network model, inputs the temporal dynamic features into the classification algorithm to realize the classification of emotions through the softmax activation function, and the emotion degree quantification unit further quantifies the classification results output by the emotion classification prediction unit into an emotion degree index to express the intensity changes of emotions.

[0020] In some possible implementations, the pre-trained questionnaire generation model includes a questionnaire rule base management unit, a dynamic question screening unit and a questionnaire structure generation unit, wherein the questionnaire rule base management unit establishes and maintains a question base containing a variety of emotional state assessment questions, wherein the question base contains basic questions, detailed assessment questions and dynamic adjustment questions for a variety of emotional states, wherein the question forms include multiple-choice questions, scale questions and open-ended questions, wherein the dynamic question screening unit is based on real-time emotional state classification and emotional degree index, and according to the question screening strategy, screens out the questions that are most suitable for the current emotional state from the question base, and the questionnaire structure generation unit dynamically generates a questionnaire structure according to the screened questions to obtain a personalized emotional assessment questionnaire, and the questionnaire structure includes question arrangement order, grouping strategy and question skipping logic.

[0021] In some possible implementations, the emotional state assessment model and questionnaire generation model are continuously optimized based on the behavioral feedback data generated by the user in the process of filling out the personalized emotional questionnaire and the physiological data collected simultaneously, including:

[0022] Obtain behavioral feedback data and synchronously collected physiological data generated by users when filling out personalized emotion assessment questionnaires;

[0023] Analyzing the behavioral feedback data and the synchronously collected physiological data to dynamically optimize question screening strategies and questionnaire structures;

[0024] Based on the accumulated questionnaire answer data and time-series dynamic feature data, the emotional state assessment model and questionnaire generation model are continuously updated;

[0025] Based on historical behavioral feedback data from similar user groups, optimize the question screening strategy and questionnaire structure for individual users.

[0026] Secondly, a personalized emotion assessment questionnaire generation system based on real-time physiological data and neural network modeling is provided, including:

[0027] A data acquisition and feature extraction module is used to acquire the user's physiological data, pre-process the physiological data and extract time series dynamic features;

[0028] An emotional state assessment module, configured to input the temporal dynamic features into a pre-trained emotional state assessment model to obtain a user's current emotional state classification and emotional degree index;

[0029] A personalized questionnaire generation module is used to input the emotional state classification and emotional degree index within a specified time period into a pre-trained questionnaire generation model to obtain a personalized emotional assessment questionnaire;

[0030] The feedback and optimization module is used to continuously optimize the emotional state assessment model and questionnaire generation model based on the behavioral feedback data generated by users when filling out personalized emotional questionnaires and the physiological data collected simultaneously.

[0031] In a third aspect, a computer program product comprising instructions is provided, which, when executed on a computer, enables the computer to execute the method described in the first aspect.

[0032] In a fourth aspect, a computer-readable storage medium is provided, wherein the computer-readable medium stores program code for execution by a device, the program code including steps for executing the method in any one of the implementations of the first aspect above.

[0033] In a fifth aspect, an electronic device is provided, comprising a processor, a memory, and a program or instruction stored in the memory and executable on the processor, wherein the program or instruction, when executed by the processor, implements a method as in any one of the implementations in the first aspect above.

[0034] The present application has the following beneficial effects: the present application realizes multimodal physiological signal fusion, high-precision emotion recognition, personalized generation and dynamic optimization of questionnaire content in the process of emotion assessment, which can effectively overcome the problems of insufficient personalization, data isolation and poor dynamic adaptability in the existing technology. It not only realizes the automatic generation of personalized emotion assessment questionnaires and improves the efficiency of emotion assessment, but also can differentiate the questionnaire content according to individual differences, thereby improving the accuracy of emotion assessment and the user experience in the process of emotion assessment. It has strong intelligence, adaptability and application promotion value, and through the feedback and optimization mechanism, it can continuously improve the accuracy of emotion assessment as data accumulates. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] The drawings that constitute a part of this application are used to provide a further understanding of this application. The illustrative embodiments of this application and their descriptions are used to explain this application and do not constitute an improper limitation on this application.

[0036] In order to more clearly illustrate the technical solutions in the embodiments of the present application, 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 application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0037] Figure 1 is a flowchart of a method for generating a personalized emotion assessment questionnaire based on real-time physiological data and neural network modeling according to Example 1 of the present application;

[0038] Figure 2 is a schematic diagram of a multimodal fusion unit in the method for generating a personalized emotion assessment questionnaire based on real-time physiological data and neural network modeling according to Example 1 of the present application;

[0039] Figure 3 This is a structural block diagram of a personalized emotion assessment questionnaire generation system based on real-time physiological data and neural network modeling according to Example 2 of the present application;

[0040] Figure 4 This is a schematic diagram of the internal structure of the electronic device of Example 4 of the present application.

[0041] Reference numerals:

[0042] 100. Data acquisition and feature extraction module; 200. Emotional state assessment module; 300. Personalized questionnaire generation module; 400. Feedback and optimization module. DETAILED DESCRIPTION

[0043] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0044] Example 1

[0045] like Figure 1 As shown, the method for generating a personalized emotion assessment questionnaire based on real-time physiological data and neural network modeling involved in Example 1 of the present application includes:

[0046] S100: Acquire multimodal physiological data of the user, preprocess the physiological data, and extract temporal dynamic features to provide basic support for subsequent emotional state assessment.

[0047] Specifically, users wear wearable devices with multimodal data collection capabilities to collect real-time heart rate (HR), heart rate variability (HRV), galvanic skin response (GSR), respiratory rate (BR), pupil dilation (PD), and electroencephalogram (EEG) signals. The acquisition frequency and accuracy are dynamically adjusted based on the signal type to ensure data quality. The collected data is then preprocessed using wavelet transforms and fast Fourier transforms (FFTs) to remove noise, smooth, and extract features from the raw physiological signals. Specific preprocessing strategies are developed for different data types to improve the effectiveness of feature extraction. The preprocessed data is then fed into a long short-term memory (LSTM) network to extract dynamic temporal features and model the user's physiological trends, providing deeper input for emotion inference.

[0048] S200: Input the temporal dynamic features into a pre-trained emotional state assessment model to obtain the user's current emotional state classification and emotional degree index, thereby realizing emotional understanding and prediction of multimodal data.

[0049] In this embodiment, the training method of the emotional state assessment model includes the following steps:

[0050] S21. Training dataset construction and annotation: Based on multi-source experimental datasets (such as DEAP and AMIGOS) and self-developed data, combined with the subjects' physiological responses and subjective self-assessment results (e.g., scale scores) in standard emotion induction tasks, expert label verification is used to annotate emotional states, including six basic emotions (joy, anger, sadness, fear, disgust, and surprise) and a neutral state. A continuous emotion level score (0–100) is also introduced.

[0051] S22. Model structure design: The backbone network is the residual network ResNet, and the cross-modal attention mechanism is integrated to improve the modeling ability of information complementarity between different signals.

[0052] S23 training and validation: Stratified sampling is used to construct the training set, validation set, and test set. The optimizer is Adam, and the loss functions include cross-entropy loss (classification) and mean square error loss (regression). The accuracy, F1 score, and RMSE are monitored during the model training process to ensure the accuracy and stability of emotion recognition, and to train an emotional state assessment model.

[0053] The pre-trained questionnaire generation model includes a questionnaire rule base management unit, a dynamic question screening unit, and a questionnaire structure generation unit, such as Figure 2As shown, the multimodal fusion unit introduces a cross-modal attention mechanism to perform weighted fusion of multimodal physiological signal features. By learning the correlations between different physiological signals, it improves the accuracy of emotional state prediction and enhances the overall model's ability to identify complex emotional states. The emotion classification prediction unit performs a comprehensive analysis of the multimodal physiological data based on a deep neural network model. The temporal dynamic features extracted in step S100 are input into the classification algorithm to classify emotions using a softmax activation function. The emotion quantification unit then quantifies the classification results output by the emotion classification prediction unit into an emotion intensity index ranging from 0 to 100 to express the intensity of the emotion. This emotion intensity index serves as a fine-grained description of emotion, helping to assess the degree of emotional fluctuation and provide personalized emotional intervention solutions. For example, the model outputs the current user's emotion classification as "anxiety" with an emotion intensity index of 75. The classification result is generated by the emotion classification prediction unit, and the emotion intensity quantification unit further outputs a quantized emotion intensity index.

[0054] S300: Input the emotional state classification and emotional level index within a specified time period into a pre-trained questionnaire generation model to obtain a personalized emotional assessment questionnaire. Filter questions through a rule library and adjust the questionnaire content in real time based on user feedback to improve the relevance and accuracy of the questionnaire.

[0055] In this embodiment, the training method of the questionnaire generation model includes the following steps:

[0056] S31. Rule library construction: Combining psychological assessment theory (such as the PHQ-9 depression scale) with actual user interview results, a standardized emotional question library is established, covering three categories: basic identification questions, detailed follow-up questions, and dynamic adaptation questions. These questions are grouped according to emotional dimensions and labeled (such as adaptive emotions, adaptation strength, semantic similarity, etc.) as the basis for screening.

[0057] S32. Dynamic Screening Logic Modeling: This approach uses a combined conditional generative model (CVAE) and decision tree modeling to generate optimal question combinations from a question library based on user-entered sentiment tags and severity indices. Training samples for this model are derived from historical user responses and simulated user profiles. The training objective is to maximize question relevance and answer acceptance rate, ultimately resulting in a questionnaire generation model.

[0058] S33. Questionnaire structure generation mechanism: In the questionnaire structure generation module, reinforcement learning strategies are applied to optimize question sorting and question skipping logic. The objective function comprehensively considers the user's time spent answering questions, the logic of the jump path, and the psychological burden index, and supports segmented question presentation and dynamic adjustment in the middle.

[0059] The pre-trained questionnaire generation model consists of a questionnaire rule library management unit, a dynamic question screening unit, and a questionnaire structure generation unit. The questionnaire rule library management unit is responsible for building and maintaining a question library containing questions assessing various emotional states. Based on psychological theories and emotional models, the question library covers basic questions, detailed assessment questions, and dynamically adjusted questions for various emotional states. Question formats include multiple-choice questions, scale questions, and open-ended questions. The dynamic question screening unit uses real-time emotion classification and emotion intensity index to select the most appropriate questions from the question library for the current emotional state. Based on emotion category and intensity, it automatically selects the most relevant positive and negative questions from the question library, ensuring that the questionnaire content is highly aligned with the user's current emotional state and avoiding interference from irrelevant questions. The questionnaire structure generation unit dynamically generates the questionnaire structure based on the question screening results. The questionnaire structure generation unit is responsible for determining the question order, grouping strategy, and question skipping logic to ensure that the questionnaire content is logical and easy to use. This flexible questionnaire structure design can effectively reduce the user's response burden. For example, the questionnaire rule library management unit matches the question item corresponding to anxiety: "Have you often felt nervous recently?" The dynamic question screening unit selects a question group that matches the current anxiety index, organizes it into a logically clear structure through the questionnaire structure generation unit, and presents it to the user.

[0060] S400: Continuously optimize the emotional state assessment model and the questionnaire generation model based on the behavioral feedback data generated by the user in the process of filling out the personalized emotional questionnaire and the physiological data collected simultaneously.

[0061] When the user fills out the personalized emotion assessment questionnaire generated in step S300, behavioral feedback data (for example, options, answering time, changes in the order of answering questions, question skipping, abandonment of questions, etc.) and synchronous physiological data (for example, heart rate variability, skin electrical response fluctuations, etc.) are collected in real time, and the above collected data are summarized to form a comprehensive feedback data set covering behavioral data and physiological data.

[0062] By analyzing the above-mentioned aggregated user feedback dataset, we dynamically optimize the question recommendation strategy and questionnaire structure, adaptively improve the evaluation accuracy and user experience in continuous interaction, and continuously update the personalized emotion recognition and questionnaire generation model based on the long-term accumulated user answer data and physiological characteristic data, so that the system can more and more accurately reflect and adapt to the user's emotional changes.

[0063] For example, if a user reacts strongly to anxiety-related questions, the reinforcement learning optimization unit dynamically adjusts the strategy to add a more specific anxiety-related question, such as "Are you afraid of the uncertainty of the future?" The user's answers in this round are recorded and correlated with their physiological data to update their emotional profile.

[0064] After comprehensive processing, an emotional state report is output, which includes: the "anxiety" emotion classification label, an emotional index of 75, the main symptom dimensions, and suggested intervention tips, namely relaxation training and psychological counseling suggestions.

[0065] In addition, based on historical feedback data from similar user groups, the question recommendations and questionnaire structure design for individual users are optimized to accelerate the formation of personalized strategies.

[0066] This embodiment provides a method for generating a personalized emotion assessment questionnaire based on real-time physiological data and neural network modeling. This method realizes multimodal physiological signal fusion, high-precision emotion recognition, personalized generation and dynamic optimization of questionnaire content in the emotion assessment process. It can effectively overcome the problems of insufficient personalization, data isolation and poor dynamic adaptability in the existing technology, improve the accuracy of emotion assessment and user experience, and has strong intelligence, adaptability and application promotion value.

[0067] Example 2

[0068] like Figure 3 As shown, the embodiment 2 of the present application involves a personalized emotion assessment questionnaire generation system based on real-time physiological data and neural network modeling, including:

[0069] The data acquisition and feature extraction module 100 is used to acquire the user's physiological data, pre-process the physiological data, and extract time series dynamic features.

[0070] Specifically, a user wears a wearable device in their daily leisure state. After the system is activated, the physiological data acquisition and modeling module monitors heart rate, galvanic skin response, and respiratory rate in real time. The collected physiological data undergoes preprocessing such as cleaning and feature extraction, and then undergoes time series modeling to form a stable physiological feature input sequence.

[0071] The emotional state assessment module 200 is used to input the temporal dynamic features into a pre-trained emotional state assessment model to obtain the user's current emotional state classification and emotional degree index.

[0072] Specifically, the pre-trained emotional state assessment model includes a multimodal fusion unit, an emotion classification prediction unit and an emotion degree quantification unit. Among them, the multimodal fusion unit introduces a cross-modal attention mechanism to perform weighted fusion of different physiological signal features to learn the correlation between different physiological signals. The emotion classification prediction unit performs a comprehensive analysis of multimodal physiological data based on a deep neural network model, inputs the time series dynamic features into the classification algorithm through the softmax activation function to realize the classification of emotions, and the emotion degree quantification unit further quantifies the classification results output by the emotion classification prediction unit into an emotion degree index to express the intensity changes of emotions.

[0073] After analyzing these features, the emotional state assessment module 200 fuses the physiological signals through the cross-modal fusion unit, and the neural network model outputs the emotion classification as "pleasure" and the emotion index as 85. This classification result is generated by the emotion classification prediction unit, and the index result is obtained by the emotion quantification unit.

[0074] The personalized questionnaire generation module 300 is used to input the emotional state classification and emotional level index within a specified time period into a pre-trained questionnaire generation model to obtain a personalized emotional assessment questionnaire.

[0075] Specifically, the pre-trained questionnaire generation model includes a questionnaire rule base management unit, a dynamic question screening unit and a questionnaire structure generation unit, wherein the questionnaire rule base management unit establishes and maintains a question base containing a variety of emotional state assessment questions, and the question base contains basic questions, detailed assessment questions and dynamic adjustment questions for a variety of emotional states, wherein the question forms include multiple-choice questions, scale questions and open-ended questions, and the dynamic question screening unit is based on real-time emotional state classification and emotional degree index, and selects the questions that are most suitable for the current emotional state from the question base according to the question screening strategy, and the questionnaire structure generation unit dynamically generates a questionnaire structure according to the screened questions to obtain a personalized emotional assessment questionnaire, and the questionnaire structure includes question arrangement order, grouping strategy and question skipping logic.

[0076] Based on the assessment results of the emotional state assessment module 200, the personalized questionnaire generation module 300 initiates the question screening process. The system's questionnaire rule library management unit retrieves positive psychological questions related to "happiness," such as "Have you often felt satisfied recently?" and "Do you enjoy communicating with others?" The dynamic question screening unit selects concise, positive questions based on the index value, and the questionnaire structure generation unit generates a lightweight questionnaire with a simple structure and a small number of questions.

[0077] The feedback and optimization module 400 is used to continuously optimize the emotional state assessment model and the questionnaire generation model based on the behavioral feedback data generated by the user in the process of filling out the personalized emotional questionnaire and the physiological data collected simultaneously.

[0078] Specifically, during the user's answering process, the system uses the user feedback collection unit of the feedback and optimization module 400 to record physiological data indicating a smooth and stable answering process and minimal fluctuations in the emotional index. Because the user's emotions are stable, the reinforcement learning optimization unit reduces the number of questions based on a reward strategy to reduce the user's burden. The collaborative filtering recommendation unit also uses recommended paths for groups with similar profiles to the current user to verify the rationality of the current questionnaire reduction strategy.

[0079] After completing the questionnaire, the system generates a personalized emotional state report, which includes the "happy" emotional label, the index value of 85, the main manifestation dimensions, and life behavior tips recommended to maintain the current good emotional state, namely "continue to maintain active social interaction", "encourage meditation or outdoor activities", etc.

[0080] In this embodiment, the data acquisition and feature extraction module 100 is used to solve the defect of data isolation in existing emotion assessment technology; the emotional state assessment module 200 and the personalized questionnaire generation module 300 are used to solve the defect of insufficient personalization; the feedback and optimization module 400 is used to solve the defect of poor dynamic adaptability. Through the collaborative work of the data acquisition and feature extraction module 100, the emotional state assessment module 200, the personalized questionnaire generation module 300 and the feedback and optimization module 400, the system realizes multimodal physiological signal fusion, high-precision emotion recognition, personalized generation and dynamic optimization of questionnaire content in the emotion assessment process, which can effectively overcome the problems of insufficient personalization, data isolation and poor dynamic adaptability in the existing technology, improve the accuracy of emotion assessment and user experience, and has strong intelligence, adaptability and application promotion value.

[0081] It should be noted that other specific implementation methods of the personalized emotion assessment questionnaire generation system based on real-time physiological data and neural network modeling in this embodiment can be found in the specific implementation methods of the personalized emotion assessment questionnaire generation method based on real-time physiological data and neural network modeling mentioned above. To avoid redundancy, they will not be repeated here.

[0082] Example 3

[0083] A computer-readable storage medium according to embodiment 3 of the present application, wherein the computer-readable storage medium stores program code for execution by a device, the program code including steps for executing the method in any one of the implementations in embodiment 1 of the present application;

[0084] Among them, the computer-readable storage medium can be a read-only memory (ROM), a static storage device, a dynamic storage device or a random access memory (RAM); the computer-readable storage medium can store program code, and when the program stored in the computer-readable storage medium is executed by the processor, the processor is used to execute the steps of the method in any one of the implementation methods in Example 1 of the present application.

[0085] Example 4

[0086] like Figure 4As shown, an electronic device involved in Example 4 of the present application includes a processor, a memory, and a program or instruction stored in the memory and executable on the processor, wherein the program or instruction, when executed by the processor, implements the method in any one of the implementations in Example 1 of the present application;

[0087] Among them, the processor can adopt a general central processing unit (CPU), a microprocessor, an application specific integrated circuit (ASIC), a graphics processing unit (GPU) or one or more integrated circuits to execute relevant programs to implement the method in any one of the implementation methods in Example 1 of the present application.

[0088] The processor may also be an integrated circuit electronic device with signal processing capabilities. In the implementation process, each step of the method in any one of the implementations in Example 1 of the present application may be completed by hardware integrated logic circuits in the processor or software instructions.

[0089] The above-mentioned processor can also be a general-purpose processor, a digital signal processor, an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic device, a discrete gate or transistor logic device, or a discrete hardware component. The various methods, steps, and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in the embodiments of the present application can be directly embodied as being executed by a hardware decoding processor, or being executed by a combination of hardware and software modules in the decoding processor. The software module can be located in a storage medium mature in the art, such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory, and the processor reads the information in the memory and, in combination with its hardware, completes the functions required to be executed by the units included in the data processing device of the embodiment of the present application, or executes the method in any one of the implementation modes in Example 1 of the present application.

[0090] The above are only preferred specific implementations of this application; however, the scope of protection of this application is not limited thereto. Any person skilled in the art who, within the technical scope disclosed in this application, makes equivalent substitutions or modifications based on the technical solutions and improved concepts of this application shall be covered by the scope of protection of this application.

Claims

1. A method for generating a personalized emotion assessment questionnaire based on real-time physiological data and neural network modeling, characterized in that: include: Acquiring multimodal physiological data of the user, and preprocessing the physiological data and extracting temporal dynamic features; Inputting the temporal dynamic features into a pre-trained emotional state assessment model to obtain the user's current emotional state classification and emotional degree index; Input the emotional state classification and emotional degree index within a specified time period into the pre-trained questionnaire generation model to obtain a personalized emotional assessment questionnaire; The emotional state assessment model and questionnaire generation model are continuously optimized based on the behavioral feedback data generated by users when filling out personalized emotional questionnaires and the physiological data collected simultaneously.

2. The method for generating a personalized emotion assessment questionnaire based on real-time physiological data and neural network modeling according to claim 1, characterized in that: The physiological data includes heart rate, heart rate variability, galvanic skin response, respiratory rate, pupil dilation degree and EEG.

3. The method for generating a personalized emotion assessment questionnaire based on real-time physiological data and neural network modeling according to claim 1, characterized in that: The preprocessing includes performing denoising, smoothing and feature extraction on the physiological signal by using wavelet transform and fast Fourier transform.

4. The method for generating a personalized emotion assessment questionnaire based on real-time physiological data and neural network modeling according to claim 1 or 3, characterized in that: The temporal feature modeling includes inputting the preprocessed physiological data into a long short-term memory network or a Transformer model to extract temporal dynamic features.

5. The method for generating a personalized emotion assessment questionnaire based on real-time physiological data and neural network modeling according to claim 4, characterized in that: The pre-trained emotional state assessment model includes a multimodal fusion unit, an emotion classification prediction unit and an emotion degree quantification unit. The multimodal fusion unit introduces a cross-modal attention mechanism to perform weighted fusion of different physiological signal features to learn the correlation between different physiological signals. The emotion classification prediction unit performs a comprehensive analysis of multimodal physiological data based on a deep neural network model, inputs the temporal dynamic features into the classification algorithm and realizes the classification of emotions through the softmax activation function. The emotion degree quantification unit further quantifies the classification results output by the emotion classification prediction unit into an emotion degree index to express the intensity changes of emotions.

6. The method for generating a personalized emotion assessment questionnaire based on real-time physiological data and neural network modeling according to claim 4, characterized in that: The pre-trained questionnaire generation model includes a questionnaire rule base management unit, a dynamic question screening unit and a questionnaire structure generation unit, wherein the questionnaire rule base management unit establishes and maintains a question base containing multiple emotional state assessment questions, wherein the question base contains basic questions, detailed assessment questions and dynamic adjustment questions for multiple emotional states, wherein the question forms include multiple-choice questions, scale questions and open-ended questions, and the dynamic question screening unit screens out the questions that are most suitable for the current emotional state from the question base according to the question screening strategy based on the real-time emotional state classification and emotional degree index, and the questionnaire structure generation unit dynamically generates a questionnaire structure according to the screened questions to obtain a personalized emotional assessment questionnaire, and the questionnaire structure includes the question arrangement order, grouping strategy and question skipping logic.

7. The method for generating a personalized emotion assessment questionnaire based on real-time physiological data and neural network modeling according to claim 6, characterized in that: The emotional state assessment model and questionnaire generation model are continuously optimized based on the behavioral feedback data generated by users when filling out personalized emotional questionnaires and the physiological data collected simultaneously, including: Obtain behavioral feedback data and synchronously collected physiological data generated by users when filling out personalized emotion assessment questionnaires; Analyzing the behavioral feedback data and the synchronously collected physiological data to dynamically optimize question screening strategies and questionnaire structures; Based on the accumulated questionnaire answer data and time-series dynamic feature data, the emotional state assessment model and questionnaire generation model are continuously updated; Based on historical behavioral feedback data from similar user groups, optimize the question screening strategy and questionnaire structure for individual users.

8. A personalized emotion assessment questionnaire generation system based on real-time physiological data and neural network modeling, characterized by: include: A data acquisition and feature extraction module is used to acquire the user's physiological data, pre-process the physiological data and extract time series dynamic features; An emotional state assessment module, configured to input the temporal dynamic features into a pre-trained emotional state assessment model to obtain the user's current emotional state classification and emotional degree index; A personalized questionnaire generation module is used to input the emotional state classification and emotional degree index within a specified time period into a pre-trained questionnaire generation model to obtain a personalized emotional assessment questionnaire; The feedback and optimization module is used to continuously optimize the emotional state assessment model and questionnaire generation model based on the behavioral feedback data generated by users when filling out personalized emotional questionnaires and the physiological data collected simultaneously.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores program codes for execution by a device, wherein the program codes include steps for executing the method according to any one of claims 1 to 7.

10. An electronic device, characterized in that: The electronic device includes a processor, a memory, and a program or instruction stored in the memory and executable on the processor, wherein the program or instruction, when executed by the processor, implements the method according to any one of claims 1 to 7.

Images