A target user health warning method based on large language model and related equipment

Abstract

The present invention provides a target user health warning method based on a large language model and related equipment, which are applied to the field of data processing technology. The present application obtains target event information, target user information, target user travel information and target user medical use information; processes the target user information based on the target event information to generate the target user's physiological identification information and the target user's attribute information; obtains an initial large language model and a training sample set that match the physiological identification information; processes the initial large language model based on the training set and the verification set to generate a target large language model; processes the target user's travel information, the target user's preset work information and the target user's preset living environment information based on the target large language model to generate the target user's physical state information; processes the target user's physical state information based on the target user's medical use information to generate warning information.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a target user health warning method based on a large language model and related equipment. Background Art

[0002] In the existing technology, the monitoring and intervention methods for the patient's inner growth mainly rely on manual assessment and standardized scales, such as the Beck Depression Inventory and the Hamilton Anxiety Rating Scale. Although these methods are widely used in clinical practice, they have many shortcomings, such as strong subjectivity, low efficiency, high resource consumption, and poor sustainability.

[0003] At present, the early identification of patients' mental disorders in clinical practice still mainly relies on manual assessment and scoring systems. Currently, the commonly used scoring methods in clinical practice include the General Health Questionnaire and the Beck Depression Inventory. Although their clinical application is very mature, they still cannot eliminate reporting bias, and respondents will intentionally or unintentionally underestimate or overestimate some information. In terms of clinical intervention, professional psychology practitioners are needed to consult and counsel patients, and provide guidance on drug use according to guidelines. This set of assessment and intervention systems has a great demand for manpower, and there is still a shortage of professionals, making it difficult to fill the gap in demand. Studies have shown that at present, there is still a lot of room for improvement in the mastery of health-related knowledge and scoring tools by medical staff in the emergency medical service system, and non-psychology medical staff still do not have the ability to provide high-quality health assessment and consulting services.

[0004] Traditional assessment methods mainly rely on interviews and observations by medical staff. The assessment results are easily affected by the subjective judgment of the assessor and have the problem of poor consistency. In addition, these methods usually require a large number of professionals, with high training costs, and it is difficult to provide continuous, real-time monitoring and timely intervention and adjustment measures.

[0005] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of the present disclosure, and therefore may include information that does not constitute the prior art known to ordinary technicians in the field. Summary of the invention

[0006] The purpose of this application is to provide a target user health warning method and related equipment based on a large language model, which at least overcomes the problems existing in the prior art to a certain extent. By analyzing the target user's response to the target event, including recall ability, emotional changes and body movements, the target user's behavior and emotional changes in the target event are monitored. By analyzing the target user's work-related actions and language, it is possible to evaluate whether the target event is related to the user's occupation, and to detect possible occupational disease risks early. Considering the impact of the target event on the user's living environment, such as business trips, moving, etc., it is possible to identify the impact of changes in the living environment on the user's physical and mental health. In addition, the introduction of non-target event information helps to more accurately judge the user's normal physiological threshold, realize personalized health monitoring and evaluation, and provide data-driven decision support for medical professionals and users based on the collected physiological and behavioral data, so as to make more targeted health intervention measures. By comprehensively analyzing the target user's physiological and psychological reactions to the target event, it can bring users beneficial effects such as comprehensive health assessment, early identification and intervention, personalized monitoring and quality of life improvement.

[0007] Other features and advantages of the present application will become apparent from the following detailed description, or may be learned in part by the practice of the present invention.

[0008] According to one aspect of the present application, a target user health warning method based on a large language model is provided, comprising: obtaining target event information, target user information, target user travel information and target user medical use information, wherein the target user information includes the target user's body movement information, the target user's facial expression information and the target user's voice information, and the target event information is information that matches the target user within a preset time period; processing the target user information based on the target event information to generate the target user's physiological identification information and the target user's attribute information, wherein the target user's physiological identification information is used to characterize the target user's mental state information within the preset time period, and the target user's attribute information is used to characterize the target user's mental state information within the preset time period. The attribute information includes the preset work information of the target user and the preset living environment information of the target user; an initial large language model and a training sample set matching the physiological identification information are obtained; the training sample set is processed based on preset processing rules to generate a training set and a verification set; the initial large language model is processed based on the training set and the verification set to generate a target large language model; the travel information of the target user, the preset work information of the target user and the preset living environment information of the target user are processed based on the target large language model to generate the physical state information of the target user; the physical state information of the target user is processed based on the medical use information of the target user to generate warning information.

[0009] In one embodiment of the present application, the target user information is processed based on the target event information to generate physiological identification information of the target user and attribute information of the target user, including: processing the target user information based on the target event information to generate abnormal action information, abnormal emotion information, sleep information and event memory information of the target user within a preset time period; processing the abnormal action information, abnormal emotion information and event memory information of the target user within a preset time period based on a preset mental health monitoring and early warning model to generate mental state information of the target user within the preset time period; generating physiological identification information of the target user based on the mental state information of the target user within the preset time period; processing the abnormal action information, abnormal emotion information and sleep information of the target user within the preset time period to generate preset work information of the target user and preset living environment information of the target user; generating attribute information of the target user based on the preset work information of the target user and the preset living environment information of the target user.

[0010] In one embodiment of the present application, the training sample set is processed based on a preset processing rule to generate a training set and a verification set, including: obtaining other user information that matches the physiological identification information of the target user; performing feature extraction on the other user information to generate a feature data set; obtaining any number of data features in the feature data set; generating adjacent features based on the distance between the any number of data features and other numbers of data features of the same category, wherein the adjacent features include a preset number of the any number of data features; determining a sampling ratio based on the number of each data feature in the training set; determining a sampling ratio based on the sampling ratio; sampling the adjacent features based on the sampling ratio to generate a preset number of sampling features; generating multiple data groups based on any data feature and each sampling feature, wherein each data group includes a preset number of data samples, wherein each data group includes a preset number of data samples, wherein at least one data sample includes physiological identification information; processing multiple data groups to generate a training set and a verification set, wherein the training set and the verification set both include a number of data features.

[0011] In one embodiment of the present application, the target user's travel information, the target user's preset work information, and the target user's preset living environment information are processed based on the target large language model to generate the target user's physical condition information, including: processing the target user's travel information and the target user's preset living environment information based on the target large language model to generate the target user's travel impact factor; processing the target user's preset work information and the target user's travel information based on the target large language model to generate the target user's work impact factor; processing the target user's travel impact factor and the target user's work impact factor to generate the target user's physical condition information.

[0012] In one embodiment of the present application, the physical state information of the target user is processed based on the medical use information of the target user to generate warning information, including: extracting features of the physical state information of the target user to generate physiological characteristic classification information, wherein the physiological characteristic classification information includes parameters corresponding to different symptom information of the target user; processing the physiological characteristic classification information based on the medical use information of the target user to generate a physiological characteristic risk influencing factor; processing the physiological characteristic risk influencing factor based on preset warning rules to generate warning information.

[0013] In one embodiment of the present application, feature extraction is performed on the target user's physical state information to generate physiological feature classification information, including: processing the target user's physical state information to generate the target user's biological sign attribute information; performing feature extraction processing on the target user's biological sign attribute information to generate a number of feature sample types; performing feature screening processing on the several feature sample types to generate real-time feature sample data; and generating physiological feature classification information based on the several real-time feature sample data.

[0014] In one embodiment of the present application, the physiological characteristic risk impact factor is processed based on a preset warning rule to generate warning information, including: obtaining the historical physiological characteristic risk impact factor of the target user, wherein the historical physiological characteristic risk impact factor of the target user is generated by the target user based on non-target event information; processing the physiological characteristic risk impact factor based on the historical physiological characteristic risk impact factor of the target user to generate a target physiological characteristic difference, wherein the physiological characteristic difference includes the sum of the differences between each physiological characteristic risk impact factor and the historical physiological characteristic risk impact factor; processing the target physiological characteristic difference based on a preset physiological characteristic threshold to generate warning information. If the target physiological characteristic difference is greater than or less than the preset physiological characteristic threshold, a warning information is generated, wherein the preset physiological characteristic threshold is generated based on the non-target event information of the target user.

[0015] Another aspect of the present application is a target user health warning device based on a large language model, characterized in that it includes: an acquisition module for acquiring target event information, target user information, target user travel information and target user medical use information, wherein the target user information includes the target user's body movement information, the target user's facial expression information and the target user's voice information, and the target event information is information that matches the target user within a preset time period; acquiring an initial large language model and a training sample set that match the physiological marker information; a processing module for processing the target user information based on the target event information to generate the target user's physiological marker information and the target user's attribute information, wherein the target user The physiological identification information of the target user is used to characterize the mental state information of the target user within a preset time period, and the attribute information of the target user includes the preset work information of the target user and the preset living environment information of the target user; the training sample set is processed based on the preset processing rules to generate a training set and a verification set; the initial large language model is processed based on the training set and the verification set to generate a target large language model; the travel information of the target user, the preset work information of the target user and the preset living environment information of the target user are processed based on the target large language model to generate the physical state information of the target user; the physical state information of the target user is processed based on the medical use information of the target user to generate warning information.

[0016] According to another aspect of the present application, an electronic device is characterized in that it includes: a first processor; and a memory for storing executable instructions of the first processor; wherein the first processor is configured to execute the above-mentioned target user health warning method based on a large language model by executing the executable instructions.

[0017] According to another aspect of the present application, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a second processor, the target user health warning method based on the large language model is implemented.

[0018] According to another aspect of the present application, a computer program product is provided, including a computer program, characterized in that when the computer program is executed by a third processor, the above-mentioned target user health warning method based on a large language model is implemented.

[0019] The application provides a target user health warning method and related equipment based on a large language model. The server analyzes the target user's response to the target event, including recall ability, emotional changes and body movements, monitors the target user's behavior and emotional changes in the target event, and analyzes the target user's work-related actions and language to evaluate whether the target event is related to the user's occupation, and to detect possible occupational disease risks as early as possible. Considering the impact of the target event on the user's living environment, such as business trips, moving, etc., it is possible to identify the impact of changes in the living environment on the user's physical and mental health. In addition, the introduction of non-target event information helps to more accurately judge the user's normal physiological threshold, realize personalized health monitoring and evaluation, and provide data-driven decision support for medical professionals and users based on the collected physiological and behavioral data, so as to make more targeted health intervention measures. By comprehensively analyzing the target user's physiological and psychological reactions to the target event, it can bring users comprehensive health assessment, early identification and intervention, personalized monitoring and quality of life improvement and other beneficial effects.

[0020] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 A flowchart of a target user health warning method based on a large language model provided by an embodiment of the present application is shown;

[0022] Figure 2 A schematic diagram of the structure of a target user health warning device based on a large language model provided in an embodiment of the present application is shown;

[0023] Figure 3 A schematic diagram of the structure of an electronic device provided by an embodiment of the present application is shown;

[0024] Figure 4 A schematic diagram of a storage medium provided in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0025] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0026] Combine the following Figure 1 To describe the target user health warning method based on a large language model according to an exemplary embodiment of the present application. It should be noted that the following application scenarios are only shown to facilitate understanding of the spirit and principles of the present application, and the embodiments of the present application are not limited in this regard. On the contrary, the embodiments of the present application can be applied to any applicable scenario.

[0027] In one implementation, the present application also proposes a target user health warning method and related equipment based on a large language model. Figure 1 The flowchart of a target user health warning method based on a large language model according to an embodiment of the present application is schematically shown. Figure 1 As shown, the method is applied to a server, comprising:

[0028] S101, obtaining target event information, target user information, target user travel information, and target user medical use information.

[0029] In one embodiment, the target user information includes the target user's body movement information, the target user's facial expression information and the target user's voice information. The target event information is information that matches the target user within a preset time period, that is, an event related to the target user occurs within a specific time period, for example, the target user currently accepts the company's arrangement to go on a business trip to another place or the target user chooses to resign and switch to another company for employment. This application does not limit the specific target event information, and the applicant can set it according to actual needs.

[0030] In addition, the target user's travel information includes but is not limited to the target user's travel mode information and the target user's destination information, among which the target user's travel information may be adjusted and changed with the target event information; the target user's medical use information mainly represents the target user's current purpose of medical treatment, including but not limited to alleviating certain current urgent symptoms (such as excessive stress in a short period of time, which requires relief and guidance; emotional frustration and self-harm due to a broken heart; excessive excitement, etc.) and solving certain long-term symptoms (such as long-term insomnia, depression, nonsense, etc.).

[0031] S102: Process the target user information based on the target event information to generate physiological identification information of the target user and attribute information of the target user.

[0032] In one embodiment, wearable physiological monitoring devices, such as smart watches and health bracelets, can be used to track personal health indicators in real time, including steps, heart rate, etc. These devices can monitor the user's movement patterns by integrating sensors, such as motion sensors, and compare them with normal patterns to identify abnormal movements. Wearable devices and health trackers can track an individual's sleep quality and record sleep cycles, depth, and number of interruptions. By analyzing this data, sleep disorders and abnormal changes in sleep patterns can be identified. Medical digital twin technology can also provide personalized health management and treatment plans by integrating multimodal data, including sleep monitoring data. The physiological signals recorded by smart wearable devices combined with artificial intelligence algorithms can be used to analyze the psychological reactions of target users to specific events. For example, by analyzing EEG data and eye movement data, changes in psychological states related to emotions and attention can be identified, thereby more accurately understanding the target user's memory and psychological reactions to specific events.

[0033] Receive the target user's voice information, facial expression information, and body movement information related to the target event information, remove noise from the collected voice data, perform word segmentation and basic semantic analysis on the denoised voice data, so as to determine the content that the target user wants to express and the memory information of the target event, evaluate the target user's cognitive function, and determine the status of his memory, attention, thinking ability, etc. Perform emotion recognition on the target user's facial expression information, identify the changes in the target user's facial expression when talking about the target event and what facial expressions the target user has, so as to identify the target user's emotional state, such as happiness, sadness, anxiety, etc. By analyzing the target user's body movements when talking about the target event, identify potential abnormal movement patterns, such as self-harm and suicidal tendencies.

[0034] The target user information is processed based on the target event information to generate the target user's abnormal action information, abnormal emotion information, sleep information and event memory information within the preset time period; the target user's abnormal action information, abnormal emotion information and event memory information within the preset time period are processed based on the preset mental health monitoring and early warning model to generate the target user's mental state information within the preset time period; the target user's physiological marker information is generated based on the target user's mental state information within the preset time period, wherein the target user's physiological marker information is used to characterize the target user's mental state information within the preset time period. That is, the target user's reaction to the target event is judged, for example, the target event's recall ability, the impact of the target event on the target user (which can be derived from the target user's actions and emotions), whether the target event is related to the target user's work (to determine whether there is a corresponding occupational disease, which can be inferred from the target user's actions and language), and when the target event occurs, whether the target user's living environment changes (for example, the need to go out, go on a business trip, move, or the current living environment has more or less other users, or the need to leave early and return late, etc., resulting in the target user's living environment being different from before).

[0035] The abnormal action information, abnormal emotion information and sleep information of the target user within a preset time period are processed to generate the preset work information of the target user and the preset living environment information of the target user; the attribute information of the target user is generated based on the preset work information of the target user and the preset living environment information of the target user, and the attribute information of the target user includes the preset work information of the target user and the preset living environment information of the target user.

[0036] Specifically, by using medical digital twin technology to analyze the target user's language, actions, and work environment, it is possible to infer whether the target event is related to the user's occupation and assess whether there is a risk of occupational disease. By analyzing the user's sleep state and facial expressions, it is possible to infer the user's psychological and emotional changes, which may be related to changes in the living environment. Studies have shown that there is a correlation between electroencephalogram (EEG) data, sleep behavior, and facial expressions during sleep, and these changes may reflect the individual's sleep quality and emotional state. For example, the study found a correlation between changes in sleep EEG data and facial expressions, among which changes in the power of the alpha band were positively correlated with sad expressions, which may suggest poor sleep quality. In addition, the frequency of sleep behavior reaches its lowest level at 3-4 hours of sleep at night, when neutral expressions appear most frequently, which may indicate that the stability of the sleep environment has a positive effect on emotions.

[0037] Furthermore, facial expression analysis technology can be used as a way to understand people's emotional state, especially in real scenes, where different facial regions play an important role in expressing different expressions. For example, the authenticity of the emotion prediction results generated by data from facial regions such as the mouth and eyes can be evaluated, which helps to understand the relationship between facial expressions and emotional states in a specific environment. The effects of acute sleep deprivation on cognitive and emotional functions have also been studied, among which damage to emotional functions is related to functional changes in brain regions such as the amygdala and medial prefrontal cortex. Sleep deprivation may cause individuals to be less sensitive to emotional valence and more likely to misjudge neutral stimuli as stimuli with higher valence. This suggests that changes in sleep status may indirectly reflect changes in the living environment by affecting emotion recognition and regulation. Based on the above information, analyzing the user's sleep status and facial expressions can provide a certain reference for inferring changes in the living environment. If a significant change in the user's sleep pattern, sleep quality, or facial expression is observed, it may be necessary to consider whether the living environment has changed and further explore the reasons behind these changes.

[0038] Under the premise of analyzing the target event, the preset work information and residence information of the target user are deduced from the action, emotion and sleep information, mainly to infer the possible work industry and living environment rather than the real and accurate information. It can effectively improve the inability to comprehensively evaluate the target user due to the target user's lack of consciousness and inability to provide accurate information. If the generated data is different from the actual situation of the target user, the preset work information and residence information can be replaced by the target user's actual work information and actual living environment information, which improves the accuracy of subsequent actual evaluation. It can more comprehensively understand the target user's response to the target event, thereby providing more accurate monitoring and intervention measures.

[0039] S103, obtaining an initial large language model and a training sample set that match the physiological marker information.

[0040] In one implementation, the initial large language model is set based on the user's physiological information received within a preset time period. That is, the initial large language model will be updated at regular intervals, because the pressure and working environment of different users will continue to change at different times. The corresponding text recognition model is generated based on the data matching the target user's time, thereby evaluating the user's mental state information within the preset time period.

[0041] S104, processing the training sample set based on preset processing rules to generate a training set and a validation set.

[0042] In one embodiment, other user information matching the physiological identification information of the target user is obtained; feature extraction is performed on the other user information to generate a feature data set; any number of data features in the feature data set are obtained; adjacent features are generated based on the distance between any number of data features and other numbers of data features of the same category, wherein the adjacent features include a preset number of any number of data features; a sampling ratio is determined based on the number of each data feature in a training set; a sampling ratio is determined based on the sampling ratio; adjacent features are sampled based on the sampling ratio to generate a preset number of sampling features; multiple data groups are generated based on any data feature and each sampling feature, wherein each data group includes a preset number of data samples, wherein each data group includes a preset number of data samples, wherein at least one data sample includes physiological identification information; multiple data groups are processed to generate a training set and a validation set, wherein both the training set and the validation set include a number of data features.

[0043] Specifically, in data analysis and machine learning, the sampling ratio and sampling rate are determined based on the number of data features in the training set. By counting the number of each data feature in the training set, the distribution of different features can be understood. The sampling ratio is determined according to the number of data features and the size of the data set. The sampling ratio can help us decide how many samples to extract from each feature. According to the sampling ratio, the specific sampling ratio can be calculated. The sampling ratio is the ratio of the actual number of samples to the total number. The adjacent features are sampled using the sampling ratio to generate a preset number of sampling features. This process ensures the representativeness and diversity of the samples. Through the sampling process, a set of sampling features is generated, which will be used for subsequent data analysis and model training. Multiple data sets are constructed based on any data feature and each sampling feature. Each data set contains a preset number of data samples. It is ensured that the samples in each data set are diverse and contain at least one sample with identification information, which helps the model learn the feature performance in different situations. These data sets are used for model training to evaluate the contribution of different features to the prediction target. The machine learning model is trained using the sampling feature set, and the performance of the model is evaluated through the validation set. According to the results of model training and validation, the sampling ratio and rate may need to be adjusted to optimize the prediction ability of the model. With this approach, large-scale data sets can be processed efficiently while maintaining data quality and model generalization capabilities, helping to identify potential security risks and abnormal behaviors.

[0044] S105 , processing the initial large language model based on the training set and the validation set to generate a target large language model.

[0045] In one embodiment, multiple data groups are extracted from a training set, wherein each data group contains a preset number of data samples, wherein at least one data sample includes physiological marker information. The initial large language model is trained based on the data samples in the multiple data groups to generate a trained large language model. In the process of developing an accurate large language model, a data group containing a preset number of samples is first extracted from the training set to ensure that each group contains at least one sample with key physiological marker information. These physiological marker information may indicate a specific physiological state or infection risk factor of the target user. Then, the initial large language model is systematically trained using these data groups to form a trained model.

[0046] The trained large language model is processed based on the validation set to generate a validation result, and then the trained model is validated using an independent validation set to generate a validation result. When analyzing the validation results, special attention is paid to samples with physiological marker information to determine whether the model can accurately identify the physiological factors that affect the user's mental state. If the model can effectively predict abnormal states, that is, samples with abnormal physiological factors, then the trained model is considered to be the target large language model. If the data sample containing physiological marker information in the validation result indicates that the physiological factors that affect the user's mental state are in an abnormal state, the trained large language model is used as the target large language model.

[0047] Use training set data to train machine learning models, such as classification models, regression models, or deep learning models, evaluate the performance of the model on the validation set, adjust model parameters and structure, and optimize them. Use the trained model to predict the risk of complications such as postoperative infection for target users, and provide doctors with personalized medical advice, such as preventive measures and treatment plans, based on the model's prediction results.

[0048] In addition, the optimization and iteration of the model is an ongoing process, which needs to be constantly adjusted according to clinical feedback and verification results to improve the accuracy and practicality of the model. After the model is applied in clinical practice, its performance is continuously monitored and regularly evaluated to ensure that the model can adapt to changes in the medical environment. Throughout the process, it is crucial to maintain data privacy and ethical standards to ensure the security and compliance of target user data. Through this process, medical professionals can build an effective early warning system to help identify infection risks in advance, optimize preventive measures, and improve the quality of postoperative recovery of target users, while reducing the burden on the medical system.

[0049] S106: Process the travel information of the target user, the preset work information of the target user, and the preset living environment information of the target user based on the target large language model to generate physical state information of the target user.

[0050] In one implementation, the target user's travel information and the target user's preset living environment information are processed based on the target large language model to generate the target user's travel influence factor. The target user's preset work information and the target user's travel information are processed based on the target large language model to generate the target user's work influence factor. The target user's travel influence factor and the target user's work influence factor are processed to generate the target user's physical state information, that is, the target user's current physical state is judged by where the target user has been, how to get to the above-mentioned places, what the purposes of each place are, the current living environment, and the user's work (which can also be the target user's actions and language).

[0051] Specifically, the target user's travel information includes travel frequency, travel mode, travel purpose, etc., as well as living environment information such as living place, living conditions, etc. The target user's preset work information includes work nature, working hours, work place, etc. The target large language model is used to extract features from the collected data to identify keywords and patterns related to travel and work. For example, features such as the frequency of travel and the choice of travel mode are extracted from the travel information. Combined with travel features and living environment features, the target large language model is used to analyze how they affect the user's physical and psychological conditions. For example, frequent travel may increase stress levels, while a good living environment may help relieve stress, or the living environment is too humid, dry, or noisy, causing sleep disorders for users. By analyzing the nature of work and travel patterns, the target large language model is used to predict how work pressure affects the user's physical state. For example, a high-intensity workload may be associated with higher psychological stress.

[0052] Combine travel influencing factors and work influencing factors, and use the target large language model to comprehensively analyze their impact on the physical state of the target user. This may include stress level, sleep quality, physical health status, etc. In addition, multimodal data such as audio, video, accelerometer data, etc. can also be combined to more comprehensively understand the behavior patterns and physical state of the target user, continuously collect the user's travel, residence and work information, and regularly update the model's training data to ensure the accuracy and timeliness of the generated physical state information, so as to effectively monitor and evaluate the user's health status.

[0053] S107: Process the target user's physical condition information based on the target user's medical use information to generate warning information.

[0054] In one embodiment, feature extraction is performed on the target user's physical state information to generate physiological feature classification information, wherein the physiological feature classification information includes parameters corresponding to different symptom information of the target user, the physiological feature classification information is processed based on the target user's medical use information to generate a physiological feature risk influencing factor, and the physiological feature risk influencing factor is processed based on preset warning rules to generate warning information.

[0055] First, it is necessary to extract key physiological features from the user's physical state information, such as heart rate, blood pressure, respiratory rate, skin conductance, etc. These data can reflect the user's emotions and stress state. For example, heart rate variability (HRV) is related to psychological stress detection. Combined with information from different sensors and data sources, such as bioinformation, language audio, facial expressions, etc., data fusion and feature extraction are performed to obtain more comprehensive and accurate mental health assessment results, taking into account individual differences, and conducting personalized assessments and interventions based on the characteristics and needs of different individuals. These features can be monitored in real time through wearable devices, and the extracted physiological features are classified according to different symptoms to form physiological feature classification information. For example, symptom parameters related to heart disease are classified into one category, symptom parameters related to breathing problems are classified into another category, and symptom parameters related to mental health are classified into one category. Specifically, for mental health, it can also be subdivided, such as the presence of neurasthenia, depression, excitement, cognitive confusion, etc.

[0056] Combined with the user's medical use information, such as medical history, drug use and treatment records, the association between physiological characteristics and mental health status is analyzed, which can help identify risk factors for specific diseases, such as depression or anxiety. By analyzing the association between physiological characteristics and medical use information, it is determined which physiological indicators may pose a risk to mental health. For example, a persistently high heart rate or blood pressure may indicate chronic stress or anxiety levels, and corresponding early warning rules are formulated. These rules can be based on specific physiological parameter thresholds, which will be expanded later. The physiological characteristic risk influencing factors are compared with the preset early warning rules. Once a high-risk signal is detected, the system will issue an early warning to prompt the user or medical professional to pay attention to potential mental health problems. It can monitor and evaluate the physiological and psychological state of the target user in real time, and provide timely early warnings and intervention suggestions when necessary.

[0057] In another embodiment, feature extraction is performed on the target user's physical state information to generate physiological feature classification information, which specifically includes the following: processing the target user's physical state information to generate the target user's biological sign attribute information, performing feature extraction processing on the target user's biological sign attribute information to generate a number of feature sample types, performing feature screening processing on the several feature sample types to generate real-time feature sample data, and generating physiological feature classification information based on the several real-time feature sample data.

[0058] Based on the needs of mental health monitoring, the target user's physical state information is processed to generate biological sign attribute information, and the process of generating physiological feature classification information through feature extraction and screening can be refined as follows: First, the target user's physiological data is collected through various sensors and monitoring devices. These data may include heart rate, blood pressure, skin conductance, breathing pattern, body movements, facial expressions, etc. These data are collectively referred to as biological sign attribute information. The collected biological sign attribute information is processed to extract features that can reflect the user's emotions and psychological state. These features may include heart rate variability (HRV), skin galvanic response, breathing frequency and depth, body movement patterns and amplitudes, changes in facial expressions, etc.

[0059] The extracted features are classified according to different emotions and psychological states to form feature sample types. For example, different feature sample types can be created according to the user's emotional state (such as calm, anxious, depressed, etc.), and the generated feature sample types are further screened to determine which features are most effective in distinguishing different psychological states. This step can use statistical methods, machine learning algorithms and other technologies to identify the most representative and discriminative features. After feature screening, a set of real-time updated feature sample data is obtained, which can reflect the changes in the psychological state of the target user in real time. Finally, based on the screened real-time feature sample data, pattern recognition and classification algorithms (such as support vector machines, random forests, deep neural networks, etc.) are used to classify the psychological state of the target user and generate physiological feature classification information. The key to this process is to accurately extract and screen features that can characterize the user's psychological state and be able to update in real time to respond to changes in the user's state. In this way, a dynamic and personalized solution can be provided for mental health monitoring.

[0060] In another embodiment, the physiological characteristic risk influence factor is processed based on a preset warning rule to generate warning information, which specifically includes the following: obtaining the historical physiological characteristic risk influence factor of the target user, wherein the historical physiological characteristic risk influence factor of the target user is generated by the target user based on non-target event information; processing the physiological characteristic risk influence factor based on the historical physiological characteristic risk influence factor of the target user to generate a target physiological characteristic difference, wherein the physiological characteristic difference includes the sum of the differences between each physiological characteristic risk influence factor and the historical physiological characteristic risk influence factor; processing the target physiological characteristic difference based on a preset physiological characteristic threshold to generate warning information, if the target physiological characteristic difference is greater than or less than the preset physiological characteristic threshold, generating warning information, wherein the preset physiological characteristic threshold is generated based on the non-target event information of the target user, that is, by introducing the non-target event information, judging what the normal threshold of the target user is, and then judging how much impact the target event information has on the target user.

[0061] Based on the acquisition and processing of risk influencing factors of historical physiological characteristics of target users, the psychological and physiological health status of users can be further analyzed and understood. The following is a detailed description of this process: First, it is necessary to collect and analyze the historical physiological data of target users under non-target events. These data may include indicators such as heart rate, blood pressure, and skin conductance. These historical data can be used as the physiological characteristic baseline of users in normal state. The physiological data of users under the influence of target events are compared with historical physiological data to identify any significant changes or deviations. This step is performed by monitoring the physiological responses of users in specific events.

[0062] Calculate the difference between each physiological characteristic risk influencing factor and the historical physiological characteristic risk influencing factor to obtain the sum of the differences. This difference reflects the degree of impact of the target event on the user's physiological state. According to the user's data under non-target events, set one or more physiological characteristic thresholds, which define the normal fluctuation range and when physiological changes may indicate potential health risks. If the target physiological characteristic difference exceeds the preset physiological characteristic threshold, the system will generate an early warning message, which indicates that the target event may have a significant negative impact on the user's physical and mental health and requires further attention or intervention. By introducing non-target event information, the user's normal physiological threshold can be judged more accurately, thereby more accurately assessing the impact of the target event. This method takes into account individual differences and improves the personalization and accuracy of the monitoring system. Effectively use physiological characteristic data to monitor and evaluate the psychological and physiological health of target users, and identify and respond to possible health risks in a timely manner.

[0063] In the present application, the server obtains target event information, target user information, travel information of the target user and medical use information of the target user, wherein the target user information includes the target user's body movement information, the target user's facial expression information and the target user's voice information, and the target event information is the information that matches the target user within a preset time period; the target user information is processed based on the target event information to generate abnormal movement information, abnormal emotion information, sleep information and event memory information of the target user within the preset time period; the abnormal movement information, abnormal emotion information and event memory information of the target user within the preset time period are processed based on the preset mental health monitoring and early warning model to generate the target user's mental state information within the preset time period; the physiological identification information of the target user is generated based on the mental state information of the target user within the preset time period; the abnormal movement information, abnormal emotion information and sleep information of the target user within the preset time period are processed to generate the target user's preset work information and the target user's preset living environment information; the attribute information of the target user is generated based on the target user's preset work information and the target user's preset living environment information, wherein the target user's physiological identification information is used to characterize the target user's mental state information within the preset time period, and the target user's attribute information includes the target user's preset work information and the target user's preset living environment information.

[0064] Acquire an initial large language model and a training sample set that match the physiological identification information; acquire other user information that matches the physiological identification information of the target user; perform feature extraction on the other user information to generate a feature data set; acquire any number of data features in the feature data set; generate adjacent features based on the distance between any number of data features and other numbers of data features of the same category, wherein the adjacent features include a preset number of any number of data features; determine a sampling ratio based on the number of each data feature in the training set; determine a sampling ratio based on the sampling ratio; sample adjacent features based on the sampling ratio to generate a preset number of sampling features; generate multiple data groups based on any data feature and each sampling feature, wherein each data group includes a preset number of data samples, wherein each data group includes a preset number of data samples, wherein at least one data sample includes physiological identification information; process the multiple data groups to generate a training set and a validation set, wherein the training set and the validation set both include a number of data features; process the initial large language model based on the training set and the validation set to generate a target large language model.

[0065] The target user's travel information and the target user's preset living environment information are processed based on the target large language model to generate the target user's travel influencing factor; the target user's preset work information and the target user's travel information are processed based on the target large language model to generate the target user's work influencing factor; the target user's travel influencing factor and the target user's work influencing factor are processed to generate the target user's physical state information; the target user's physical state information is processed to generate the target user's biological sign attribute information; feature extraction processing is performed on the target user's biological sign attribute information to generate a number of feature sample types; feature screening processing is performed on a number of feature sample types to generate real-time feature sample data; physiological feature classification information is generated based on a number of real-time feature sample data, wherein the physiological feature classification information includes parameters corresponding to different symptom information of the target user; the physiological feature classification information is processed based on the target user's medical use information to generate a physiological feature risk influencing factor.

[0066] In addition, the historical physiological characteristic risk influencing factors of the target user will be obtained, wherein the historical physiological characteristic risk influencing factors of the target user are generated by the target user based on non-target event information; the physiological characteristic risk influencing factors are processed based on the historical physiological characteristic risk influencing factors of the target user to generate a target physiological characteristic difference, wherein the physiological characteristic difference includes the sum of the differences between each physiological characteristic risk influencing factor and the historical physiological characteristic risk influencing factor; the target physiological characteristic difference is processed based on the preset physiological characteristic threshold to generate warning information. If the target physiological characteristic difference is greater than or less than the preset physiological characteristic threshold, a warning information is generated, wherein the preset physiological characteristic threshold is generated based on the non-target event information of the target user. By analyzing the target user's response to the target event, including recall ability, emotional changes and body movements, monitoring the target user's behavior and emotional changes in the target event, and by analyzing the target user's work-related movements and language, it is possible to evaluate whether the target event is related to the user's occupation, and to discover possible occupational disease risks early. Considering the impact of the target event on the user's living environment, such as business trips, moving, etc., it is possible to identify the possible impact of changes in the living environment on the user's physical and mental health. In addition, the introduction of non-target event information can help to more accurately determine the user's normal physiological threshold, realize personalized health monitoring and evaluation, and provide data-driven decision support for medical professionals and users based on the collected physiological and behavioral data, so as to make more targeted health intervention measures. By comprehensively analyzing the physiological and psychological reactions of target users to target events, it can bring users beneficial effects such as comprehensive health assessment, early identification and intervention, personalized monitoring and improvement of quality of life.

[0067] In one embodiment, if Figure 2 As shown, the present application also provides a target user health warning device based on a large language model, comprising:

[0068] The acquisition module 201 is used to acquire target event information, target user information, travel information of the target user and medical use information of the target user, wherein the target user information includes body movement information of the target user, facial expression information of the target user and voice information of the target user, and the target event information is information matching the target user within a preset time period; and to acquire an initial large language model and a training sample set matching the physiological marker information;

[0069] The processing module 202 is used to process the target user information based on the target event information to generate physiological identification information of the target user and attribute information of the target user, wherein the physiological identification information of the target user is used to characterize the mental state information of the target user within a preset time period, and the attribute information of the target user includes the preset work information of the target user and the preset living environment information of the target user; process the training sample set based on preset processing rules to generate a training set and a verification set; process the initial large language model based on the training set and the verification set to generate a target large language model; process the travel information of the target user, the preset work information of the target user and the preset living environment information of the target user based on the target large language model to generate the physical state information of the target user; process the physical state information of the target user based on the medical use information of the target user to generate warning information.

[0070] The present application embodiment provides an electronic device, such as Figure 3 As shown, the electronic device 3 includes a first processor 300, a memory 301, a bus 302 and a communication interface 303, and the first processor 300, the communication interface 303 and the memory 301 are connected through the bus 302; the memory 301 stores a computer program that can be run on the first processor 300, and when the first processor 300 runs the computer program, the target user health warning method based on the large language model provided in any of the aforementioned embodiments of the present application is executed.

[0071] The memory 301 may include a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk memory. The communication connection between the system network element and at least one other network element is realized through at least one communication interface 303 (which may be wired or wireless), and the Internet, wide area network, local area network, metropolitan area network, etc. may be used.

[0072] The bus 302 may be an ISA bus, a PCI bus, or an EISA bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. The memory 301 is used to store a program, and the first processor 300 executes the program after receiving an execution instruction. The target user health warning method based on a large language model disclosed in any implementation of the embodiment of the present application may be applied to the first processor 300, or implemented by the first processor 300.

[0073] The first processor 300 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit or software instructions in the first processor 300. The above-mentioned first processor 300 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a readily available programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The 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 embodied as a hardware decoding processor to execute, or a combination of hardware and software modules in the decoding processor to execute. The software module can be located in a mature storage medium in the field 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 301, and the first processor 300 reads the information in the memory 301 and completes the steps of the above method in combination with its hardware.

[0074] The electronic device provided by the above-mentioned embodiments of the present application and the target user health warning method based on a large language model provided by the embodiments of the present application are based on the same inventive concept and have the same beneficial effects as the methods adopted, run or implemented by the applications stored therein.

[0075] The present application embodiment provides a computer-readable storage medium, such as Figure 4 As shown, the computer-readable storage medium stores 401 a computer program, and when the computer program is read and executed by the second processor 402, the target user health warning method based on the large language model as described above is implemented.

[0076] The technical solution of the embodiment of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for enabling an electronic device (which may be an air conditioner, a refrigeration device, a personal computer, a server, or a network device, etc.) or a processor to execute all or part of the steps of the method described in the embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a ROM, a RAM, a magnetic disk, or an optical disk.

[0077] The computer-readable storage medium provided in the above-mentioned embodiments of the present application and the target user health warning method based on a large language model provided in the embodiments of the present application are based on the same inventive concept and have the same beneficial effects as the methods adopted, run or implemented by the application programs stored therein.

[0078] An embodiment of the present application provides a computer program product, including a computer program, wherein the computer program is executed by a third processor to implement the method described above.

[0079] The computer program product provided by the above-mentioned embodiments of the present application and the target user health warning method based on a large language model provided by the embodiments of the present application are based on the same inventive concept and have the same beneficial effects as the methods adopted, run or implemented by the application programs stored therein.

[0080] It should be noted that, in this application, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or still includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise one..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0081] Each embodiment in the present application is described in a related manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for evaluating the target user health warning method based on a large language model, electronic device, electronic device, and readable storage medium embodiment, since it is basically similar to the target user health warning method embodiment based on a large language model described above, the description is relatively simple, and the relevant parts can be referred to the partial description of the target user health warning method embodiment based on a large language model described above.

[0082] Although the present application is disclosed as above, the present application is not limited thereto. Any person skilled in the art can make various changes and modifications without departing from the spirit and scope of the present application. Therefore, the protection scope of the present application shall be subject to the scope defined by the claims.

Claims

1. A target user health warning method based on a large language model, characterized in that: include: Acquire target event information, target user information, target user travel information and target user medical use information, wherein the target user information includes target user body movement information, target user facial expression information and target user voice information, and the target event information is information that matches the target user within a preset time period; Processing the target user information based on the target event information to generate physiological identification information of the target user and attribute information of the target user, wherein the physiological identification information of the target user is used to characterize the mental state information of the target user within a preset time period, and the attribute information of the target user includes preset work information of the target user and preset living environment information of the target user; Acquire an initial large language model and a training sample set that match the physiological marker information; Processing the training sample set based on preset processing rules to generate a training set and a validation set; Processing the initial large language model based on the training set and the validation set to generate a target large language model; Processing the target user's travel information, the target user's preset work information, and the target user's preset living environment information based on the target large language model to generate the target user's physical condition information; The physical condition information of the target user is processed based on the medical use information of the target user to generate warning information.

2. The method according to claim 1, characterized in that The target user information is processed based on the target event information to generate physiological identification information of the target user and attribute information of the target user, including: Processing the target user information based on the target event information to generate abnormal action information, abnormal emotion information, sleep information and event memory information of the target user within a preset time period; Based on the preset mental health monitoring and early warning model, the abnormal action information, abnormal emotion information and event memory information of the target user within the preset time period are processed to generate the mental state information of the target user within the preset time period; Generate physiological identification information of the target user based on the mental state information of the target user within a preset time period; Processing the target user's abnormal action information, abnormal emotion information, and sleep information within a preset time period to generate the target user's preset work information and the target user's preset living environment information; The attribute information of the target user is generated based on the preset work information of the target user and the preset living environment information of the target user.

3. The method according to claim 1, characterized in that The training sample set is processed based on a preset processing rule to generate a training set and a validation set, including: Acquire other user information that matches the physiological identification information of the target user; Extract features from other user information and generate feature data sets; Obtaining any number of data features in the feature data set; Generate adjacent features based on the distance between the any number of data features and other numbers of data features of the same category, wherein the adjacent features include a preset number of the any number of data features; Determining a sampling ratio based on the number of each data feature in the training set; Based on the sampling ratio, determining a sampling ratio; Sampling the adjacent features based on the sampling ratio to generate a preset number of sampling features; Based on any data feature and each sampling feature, generating a plurality of data groups, wherein each data group includes a preset number of data samples, and at least one data sample includes physiological identification information; The plurality of data sets are processed to generate a training set and a validation set, wherein the training set and the validation set both include a number of data features.

4. The method according to claim 2, characterized in that The target user's travel information, the target user's preset work information, and the target user's preset living environment information are processed based on the target large language model to generate the target user's physical state information, including: Processing the travel information of the target user and the preset living environment information of the target user based on the target large language model to generate a travel influencing factor of the target user; Processing the preset work information of the target user and the travel information of the target user based on the target large language model to generate a work influence factor of the target user; The travel influencing factor of the target user and the work influencing factor of the target user are processed to generate physical status information of the target user.

5. The method according to claim 4, characterized in that Processing the target user's physical condition information based on the target user's medical use information to generate warning information includes: Extracting features of the target user's physical state information to generate physiological feature classification information, wherein the physiological feature classification information includes parameters corresponding to different symptom information of the target user; Processing the physiological characteristic classification information based on the medical use information of the target user to generate a physiological characteristic risk influencing factor; The physiological characteristic risk influencing factors are processed based on preset warning rules to generate warning information.

6. The method according to claim 5, characterized in that Extracting features from the target user's physical state information to generate physiological feature classification information includes: Processing the target user's physical condition information to generate the target user's biological attribute information; Performing feature extraction processing on the biological attribute information of the target user to generate a plurality of feature sample types; Performing feature screening processing on the several feature sample types to generate real-time feature sample data; Generate physiological feature classification information based on a number of real-time feature sample data.

7. The method according to claim 5, characterized in that The physiological characteristic risk influencing factors are processed based on preset warning rules to generate warning information, including: Acquire a historical physiological characteristic risk impact factor of a target user, wherein the historical physiological characteristic risk impact factor of the target user is generated by the target user based on non-target event information; Processing the physiological characteristic risk impact factor based on the historical physiological characteristic risk impact factor of the target user to generate a target physiological characteristic difference, wherein the physiological characteristic difference includes the sum of the differences between each physiological characteristic risk impact factor and the historical physiological characteristic risk impact factor; If the target physiological characteristic difference is greater than or less than a preset physiological characteristic threshold, a warning message is generated, wherein the preset physiological characteristic threshold is generated based on non-target event information of the target user.

8. A target user health warning device based on a large language model, characterized in that: The device comprises: An acquisition module is used to acquire target event information, target user information, target user travel information and target user medical use information, wherein the target user information includes target user body movement information, target user facial expression information and target user voice information, and the target event information is information that matches the target user within a preset time period; acquire an initial large language model and a training sample set that match the physiological marker information; A processing module is used to process the target user information based on the target event information to generate physiological identification information of the target user and attribute information of the target user, wherein the physiological identification information of the target user is used to characterize the mental state information of the target user within a preset time period, and the attribute information of the target user includes the preset work information of the target user and the preset living environment information of the target user; process the training sample set based on preset processing rules to generate a training set and a verification set; process the initial large language model based on the training set and the verification set to generate a target large language model; process the travel information of the target user, the preset work information of the target user and the preset living environment information of the target user based on the target large language model to generate the physical state information of the target user; process the physical state information of the target user based on the medical use information of the target user to generate warning information.

9. An electronic device, characterized in that: include: a first processor; and a memory for storing executable instructions of the first processor; The first processor is configured to execute the target user health warning method based on a large language model as described in any one of claims 1 to 7 by executing the executable instructions.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the second processor, the target user health warning method based on the large language model described in any one of claims 1 to 7 is implemented.

Images