Management method for automatically monitoring user and related equipment

Abstract

The invention provides a management method for automatically monitoring a user and related equipment, and is applied to the technical field of data processing. The method comprises the following steps: processing biological characteristic information of a target user to generate identity attribute information of the target user; processing the medical record information of the target user and the physiological type information of the target user to generate user rating information; preprocessing the training sample set to generate a training sample set with target feature data; processing a preset user information screening model based on the training sample set with the target feature data to generate a target user information screening model; processing the user rating information and the identity information of the target user based on a target user information screening model to generate initial monitoring information of the target user; and processing the initial monitoring information of the target user based on the consultation purpose information of the target user and the emotion information of the target user, and generating target monitoring information of the target user.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a management method and related equipment for automatically monitoring users. Background Art

[0002] Clinical care is based on information that is shared across multiple care roles to coordinate care, including physicians, nurses, users, family, etc. In many settings today, there is an overload of care information that may not be known to all team members, may not be relevant at a given decision point, or may be relevant at the time of the decision but unknown to the decision-making member.

[0003] In many hospitals, there are many users but few pharmacists. Pharmacists can only perform key monitoring on some users. If the screening is inaccurate, some users who need key monitoring cannot get effective monitoring services, which will reduce user satisfaction. Existing medical monitors usually push alarm messages to the central station. However, there are many medical monitors connected to the central station in the department. Every time the central station receives an alarm message, medical staff need to check whether it is an alarm message of the user they are concerned about. When there are many alarm messages, medical staff need to spend more time to check the alarm messages of the users they are concerned about, and it is easy to miss them.

[0004] 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

[0005] The purpose of this application is to provide a management method and related equipment for automatic monitoring of users, at least to a certain extent, to overcome the problems existing in the prior art, by focusing on the automatic monitoring management of users, and obtaining multiple information, including biometrics, medical records, consultation purpose information, as well as preset models and training sample sets. The biometric information is processed in series to generate identity attribute information, and the user rating information is obtained after the medical records and physiological type information are processed. The training sample set processing is used to construct a target user information screening model. Then, based on the screening model and related information, real-time monitoring level information is generated, and the initial monitoring information is obtained after determining the monitoring priority. Finally, classification, early warning and dynamic adjustment information are generated based on the consultation purpose and emotional information, and the initial monitoring information is optimized to the target monitoring information, so as to achieve accurate adjustment of the monitoring plan. The entire process covers multi-faceted data processing and information integration, aiming to provide users with accurate, dynamic and effective automatic monitoring services, improve the quality and efficiency of medical monitoring, and meet the medical needs of different users.

[0006] 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.

[0007] According to one aspect of the present application, a management method for automatically monitoring a user is provided, comprising: obtaining biometric information of a target user, medical history information of the target user, consultation purpose information of the target user, a preset user information screening model and a training sample set; processing the biometric information of the target user to generate identity attribute information of the target user, wherein the identity attribute information of the target user includes identity information of the target user, physiological type information of the target user and emotional information of the target user; processing the medical history information of the target user and the physiological type information of the target user to generate user rating information; preprocessing the training sample set to generate a training sample set with target feature data, wherein the target feature data is used to characterize risk factors affecting the physiological characteristics of the user; processing the preset user information screening model based on the training sample set with target feature data to generate a target user information screening model; processing the user rating information and the identity information of the target user based on the target user information screening model to generate initial monitoring information of the target user; processing the initial monitoring information of the target user based on the consultation purpose information of the target user and the emotional information of the target user to generate target monitoring information of the target user.

[0008] Another aspect of the present application is a management device for automatically monitoring a user, characterized in that it includes: an acquisition module, which is used to acquire the biometric information of a target user, the medical history information of the target user, the consultation purpose information of the target user, a preset user information screening model and a training sample set; a processing module, which is used to process the biometric information of the target user to generate the identity attribute information of the target user, wherein the identity attribute information of the target user includes the identity information of the target user, the physiological type information of the target user and the emotional information of the target user; the medical history information of the target user and the physiological type information of the target user are processed to generate user rating information; the training sample set is pre-processed to generate a training sample set with target feature data, wherein the target feature data is used to characterize the risk factors affecting the physiological characteristics of the user; based on the training sample set with the target feature data, the preset user information screening model is processed to generate a target user information screening model; based on the target user information screening model, the user rating information and the identity information of the target user are processed to generate the initial monitoring information of the target user; based on the consultation purpose information of the target user and the emotional information of the target user, the initial monitoring information of the target user is processed to generate the target monitoring information of the target user.

[0009] 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 above-mentioned management method for automatically monitoring a user is implemented.

[0010] The present application provides a management method and related equipment for automatic monitoring of users, focusing on automatic monitoring management of users, with diversified information acquisition, including biometrics, medical records, consultation purpose information, as well as preset models and training sample sets. The biometric information is processed in series to generate identity attribute information, and the user rating information is obtained after the medical records and physiological type information are processed. The training sample set processing is used to construct a target user information screening model. Next, real-time monitoring level information is generated based on the screening model and related information, and the initial monitoring information is obtained after determining the monitoring priority. Finally, classification, early warning and dynamic adjustment information are generated based on the consultation purpose and emotional information, and the initial monitoring information is optimized to the target monitoring information, so as to achieve precise adjustment of the monitoring plan. The entire process covers multi-faceted data processing and information integration, aiming to provide users with accurate, dynamic and effective automatic monitoring services, improve the quality and efficiency of medical monitoring, and meet the medical needs of different users.

[0011] 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

[0012] Figure 1 A flowchart showing a method for automatically monitoring a user provided by an embodiment of the present application is shown;

[0013] Figure 2 A schematic diagram of the structure of a management device for automatically monitoring a user provided by an embodiment of the present application is shown. DETAILED DESCRIPTION

[0014] 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.

[0015] Combine the following Figure 1 To describe the management method of automatically monitoring a user 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 principle 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.

[0016] In one implementation, the present application also proposes a management method and related equipment for automatically monitoring users. Figure 1The flowchart of a method for automatically monitoring a user according to an embodiment of the present application is schematically shown. Figure 1 As shown, the method is applied to a server, comprising:

[0017] S101, obtaining the target user's biometric information, the target user's medical record information, the target user's consultation purpose information, a preset user information screening model and a training sample set.

[0018] In one implementation, high-definition cameras are installed at various entrances of the hospital (such as the outpatient hall, the gate of the inpatient department, etc.) and key areas (such as the pharmacy, the examination department, etc.) to collect facial images of target users who come for medical treatment or handle related business. During the collection process, ensure that there is sufficient light and the image is clear so as to accurately obtain information such as facial features and facial contours. These images are not only used for identity recognition, but also provide basic data for subsequent emotion analysis. For example, when a user enters the outpatient hall, the camera automatically captures his or her front-facing bareheaded image, stores it in the system, and associates it with the user's identity information.

[0019] In some areas with high requirements for identity authentication, such as the collection of special drugs and the use of high-value medical equipment, fingerprint recognition devices are set up. When the target user performs relevant operations, he needs to place his finger on the fingerprint recognition device for fingerprint collection. The collected fingerprint information is encrypted and stored, which is used to uniquely identify the user's identity and perform identity authentication in subsequent business processes. For example, when a user collects special controlled drugs, he first confirms his identity through fingerprint recognition to ensure the safety and accuracy of drug distribution. At the same time, the fingerprint information also becomes part of his biometric information. When the target user interacts with the hospital's intelligent customer service system (such as through telephone consultation, online voice consultation, etc.), the system automatically records his voice information. During the voice collection process, a high-sensitivity microphone is used to ensure the quality of the voice signal. The collected voice features include pitch, speech speed, timbre, voice rhythm, etc. These features can reflect some of the user's physiological and psychological states. For example, when a user calls the hospital consultation hotline, the system silently records his voice features while answering the user's questions, providing auxiliary data for subsequent emotional analysis or user feature identification.

[0020] Obtain the target user's basic medical record information from the hospital's information management system (such as the HIS system), including name, age, gender, contact information, home address, etc. This information helps to identify the user and understand the basic situation, facilitate communication and contact with the user in subsequent services, and can also be used as basic data for user classification management. For example, based on age and gender information, health service demand analysis can be conducted for user groups of different age groups and genders. Query the target user's previous medical records in the hospital, mainly to obtain information such as their visit time, visit department, and visit purpose (such as routine physical examination, preventive health consultation). These records can reflect the user's usage habits and demand tendencies for medical services, and provide a reference for hospitals to optimize service processes and reasonably allocate medical resources. For example, statistics are collected on the user's visit frequency in different time periods, as well as the degree of preference for services in different departments, so that the hospital can make corresponding service preparations in advance. Collect the target user's inspection and testing records, including the name of the inspection and testing item, the inspection and testing time, and the inspection and testing agency (if an external hospital inspection is involved). These records can help understand users' awareness of health management and their participation in different examination and testing items. For example, they can count the proportion of users participating in various routine physical examination items (such as blood routine, urine routine, liver function tests, etc., and analyze whether the test results are abnormal), providing a basis for hospitals to carry out health education and promote related examination and testing services.

[0021] Analyze the consultation records of target users on the hospital's online consultation platform, focusing on the subject classification of their consultations (such as medical service process consultation, hospital facility use consultation, medical insurance policy consultation, etc.), consultation time, consultation method (text consultation, voice consultation, etc.) and other information. For example, count the number of users' consultations on medical insurance reimbursement processes in different time periods to understand the users' doubts about medical insurance policies, so that hospitals can optimize medical insurance service consultations and better answer medical insurance-related questions for users. When the target user calls the hospital's customer service hotline, the customer service staff will record the content of their consultations in detail, such as hospital appointment registration rules, outpatient department distribution, inpatient environment facilities and other issues. At the same time, record the duration of the consultation, the user's tone and attitude (used to evaluate user satisfaction) and other information. For example, by analyzing the types and frequency of questions that users consult about the hospital's appointment registration process, it is found that users are prone to misunderstandings in a certain appointment link, so as to optimize the guidance instructions for appointment registration in a targeted manner and improve user experience. At various consultation service points in the hospital (such as the consultation desk in the outpatient hall, the nurse station in the inpatient department, etc.), staff will register the on-site consultations of target users. The registration content mainly includes non-medical advice questions (such as transportation information around the hospital, catering and accommodation recommendations, etc.), consultation location, consultant identity type (user himself, family members, accompanying personnel), etc. For example, based on the on-site consultation registration information, the hospital can learn about the needs of the user's family members for surrounding living facilities during the user's hospitalization. The hospital can cooperate with surrounding businesses to provide users and their families with more convenient life service guides.

[0022] The preset user information screening model is a model based on machine learning algorithms, which collects a large amount of non-disease-related behavior and feature data about users in the hospital, such as the user's activity trajectory in the hospital (recorded by the hospital's positioning system or surveillance cameras), browsing history on the hospital website or mobile app (such as browsing hospital news, online reservation function usage habits, etc.), evaluation feedback on hospital services, etc. These data are cleaned and preprocessed to remove invalid data and outliers, and the data is converted into a format suitable for model training, such as segmenting and vectorizing text data, and normalizing numerical data.

[0023] Select features related to user service needs and behavior patterns from the preprocessed data, such as the average length of stay of users in the hospital, the frequency of use of different hospital service modules (such as registration services, payment services, report query services, etc.), and the degree of attention paid to non-medical information content (such as hospital culture propaganda, health activity promotion, etc.) on hospital websites or applications. Use these features to train machine learning models, such as clustering models (classifying users according to their behavior patterns, such as active users, ordinary users, etc., but not based on disease factors) and association rule mining models (discovering the association between user behaviors, such as whether users who frequently use online appointment functions are also more inclined to use electronic report query functions). Use common evaluation indicators (such as accuracy and recall for classification models, mean square error for regression models, etc., but not involving disease-related evaluation standards) to evaluate the trained model, and adjust model parameters, improve feature selection methods, or try different algorithms based on the evaluation results to improve model performance. For example, by adjusting the number of clusters in the clustering model, observe whether the accuracy of the model in classifying user behavior is improved, and optimize the model's ability to predict user service needs.

[0024] Collect relevant data of target users from various information systems of the hospital as part of the training sample set, such as user consumption records in the hospital (including catering consumption, parking payment, convenience store shopping, etc.), records of participating in health lectures or activities held by the hospital (such as participating in healthy lifestyle lectures, volunteer activities, etc.), feedback and evaluation of hospital environment facilities (such as evaluation of ward comfort, hospital green environment, etc.). These data can reflect the comprehensive experience and demand tendency of users in the hospital and provide data support for building user behavior models. Record the non-medical behavior data of target users in the hospital, such as the stay time and activity trajectory in the public areas of the hospital (waiting areas, corridors, etc.), the interaction with hospital staff (such as asking for directions, seeking help, etc.), the frequency and time of using hospital public facilities (such as water dispensers, charging stations, etc.). These behavioral data can help analyze the daily behavior patterns of users in the hospital, provide a basis for optimizing the layout of hospital service facilities and service processes, and at the same time, as part of the training sample set, it is used to train models related to user behavior pattern recognition.

[0025] S102: Process the biometric information of the target user to generate identity attribute information of the target user.

[0026] In one embodiment, the biometric information of the target user is processed to generate initial facial image information, identity information of the target user, and physiological type information of the target user. The collected facial images are transmitted to the background image processing system in real time. The system first screens the images to remove images with poor quality due to dark light, occlusion, etc. Then, the retained images are preprocessed, including grayscale processing to simplify subsequent calculations; image noise reduction to remove noise interference caused by equipment or environmental factors; and image enhancement operations, such as histogram equalization, to improve the contrast and clarity of the image. The preprocessed image is feature extracted using an advanced face recognition algorithm. The algorithm accurately locates key facial feature points, such as the inner and outer corners of the eyes, the center of the pupil, the tip and wings of the nose, the corners of the mouth and the lip contours. A geometric model of the face is constructed through these feature points to generate initial facial image information, which includes data such as the basic shape of the face, the position of the facial features, and the relative proportions. This application does not limit the specific algorithm, and the applicant can choose according to actual needs. When collecting biometric information, users need to provide valid identification documents (such as ID cards, medical insurance cards, social security cards, etc.) or existing registration information in the hospital information system. The system uses optical character recognition (OCR) technology or matches with database information to extract the user's identity information, including name, date of birth (for calculating age), gender, ID number, contact number, etc. The extracted identity information is associated and bound with the initial facial image information just generated to create a unique user record in the database. For example, using the ID number as the unique identifier, the user's facial image information, identity information, etc. are stored in the same data entry to facilitate subsequent rapid query and identification.

[0027] The user's age is calculated based on the date of birth in the acquired identity information, and the physiological type is preliminarily divided according to the age range. For example, users under the age of 18 are classified as minors; users aged 60 or above are classified as elderly people; and those between 18 and 60 are adults. At the same time, gender information is also an important part of the physiological type and is used for subsequent analysis (different genders have differences in physiological characteristics and health risks, which can serve as the basis for subsequent analysis, such as the distribution patterns of certain health conditions in different genders).

[0028] The initial face image information is processed based on the preset processing rules to generate the influencing factors of the target user's expression characteristics. The facial area in the initial face image information is divided into several sub-areas that are closely related to the expression of expression. It mainly includes the eye area, which is defined as a specific rectangular range centered on the eyes; the mouth area, which covers the range of muscle activity of the lips and its surroundings; and the eyebrow area, which is the narrow area where the eyebrows are located. For each sub-area, a detailed feature analysis method is formulated. For the eye area, the focus is on the degree of eye opening, eyelid movement (such as blinking frequency, blinking duration), changes in eye wrinkles (such as the appearance and deepening of crow's feet) and other features; the mouth area focuses on analyzing the shape changes of the lips (such as pursing, grinning, pouting, etc.), the angle of the corners of the mouth, the degree of opening and closing of the lips, etc.; the eyebrow area mainly observes the height of the eyebrows, the change in the distance between the eyebrows (such as the distance becomes narrower when frowning), and the tilt direction of the eyebrows. The characteristic changes in these areas together constitute the influencing factors of expression characteristics.

[0029] The system has the ability to process and analyze video streams in real time, and can dynamically track continuous facial image frames. Based on the initial facial image information, it marks and tracks the position changes and movement trajectories of key facial expression feature points (such as the feature points in the eye, mouth, and eyebrow areas mentioned above) between different frames. For example, it records the number of eye blinks in a period of time, the duration of each blink, and the change in the direction of eye gaze; the opening and closing speed of the mouth during speaking or expression changes, the maximum opening angle, and the direction and amplitude of the movement of the mouth corners; the speed of eyebrows rising and falling when the expression changes, and the time they maintain a certain shape. These dynamic tracking data can more comprehensively reflect the changing process and trend of the user's expression, and provide richer information for accurately judging the factors affecting the expression characteristics (the dynamic characteristics of certain health conditions that affect facial expressions, such as abnormal facial muscle movements caused by neurological problems, can be used as the direction of subsequent potential analysis).

[0030] Combined with the environmental information when collecting facial images, such as ambient light intensity, temperature, humidity, etc. (obtained through sensors installed at corresponding locations), the potential impact of environmental factors on user expressions is analyzed. For example, under strong light, users will unconsciously squint their eyes. This change in expression is mainly caused by the ambient light, not the internal emotional expression; in a high temperature environment, users will experience facial skin flushing and forehead sweating due to the heat. These physiological reactions will affect the presentation of facial expressions, such as causing eyebrows to wrinkle slightly and mouth to open slightly. By establishing an association model or knowledge base between environmental factors and changes in expression characteristics, the system can better identify and distinguish which expression changes are caused by environmental factors and which ones truly reflect the user's internal emotions or other factors when analyzing the influencing factors of expression characteristics. This can improve the accuracy of the judgment of the influencing factors of expression characteristics and avoid misjudgment due to interference from environmental factors (the impact of environmental factors on the physiological state of the human body is indirectly related to the manifestation of certain health problems, which can be used as clues for subsequent in-depth analysis).

[0031] The influencing factors of the target user's expression characteristics are processed to generate target face image information, wherein the target face image information includes the expression change information of the target user. The influencing factors of the expression characteristics obtained through the above processing are quantified, and the characteristic changes of each facial sub-region are converted into a calculable value. For example, for the degree of eye opening, the ratio of the height of the open part of the eye to the height of the entire eye region can be calculated, and the value range is set to 0 (completely closed) to 1 (completely open); for the angle of the mouth corner, it is represented by the angle of the mouth corner relative to the initial horizontal position, and the angle range can be set to -90 degrees (maximum droop) to 90 degrees (maximum upward); for the eyebrow raising height, the vertical displacement of the upper edge of the eyebrow relative to the initial position is measured, and normalized according to the facial size to obtain a relative value. The quantized expression characteristic influencing factors are encoded and converted into a format suitable for computer storage and processing. A common method is to use vector encoding to combine the quantized characteristic values ​​of each sub-region such as the eye, mouth, and eyebrow into a characteristic vector in a certain order. For example, a simple feature vector can be expressed as [eye openness value, mouth corner upturn angle value, eyebrow upturn height value]. Through this encoding method, each specific expression state can be represented by a unique feature vector, thereby generating a digital form of the target facial image information containing rich expression change information.

[0032] The target facial image information is identified and classified by the expression change patterns using the pre-trained expression classification model. These models can be built based on machine learning algorithms (such as support vector machines, decision trees, etc.) or deep learning architectures (such as convolutional neural networks), and the training data is a large number of facial image samples labeled with different expression states. The model matches and compares the feature vector of the input target facial image information with the expression pattern in the training data, and determines whether the expression belongs to a common expression category, such as joy), sadness, anger, surprise, fear, disgust, etc., or some intermediate state or complex mixed expression between these basic expressions. For example, if the feature vector is highly similar to the feature pattern of the joy expression in the training data, the model determines that the user's current expression is of the joy type. This recognition result will provide an important basis for the subsequent generation of emotional information (different expression patterns are related to certain potential health conditions, such as long-term abnormal expressions suggesting psychological or physiological problems, which can be used as a subsequent research direction).

[0033] Process the target face image information to generate the target user's emotional information. Establish a mapping relationship between expression features and emotions. For example, in general, expression feature combinations such as eyes wide open, eyebrows raised, and mouth corners raised are often associated with positive emotional states (similar to emotions such as happiness and excitement); while expression feature combinations such as frowning, drooping mouth corners, and squinting eyes usually correspond to negative emotional states (similar to emotions such as sadness and dissatisfaction). Organize the mapping relationship between these expression features and emotions into a knowledge base or database, which records the emotional ranges corresponding to various expression feature combinations and the probability distribution relationship between them. For example, in this knowledge base, for the expression feature combination of eyes wide open, eyebrows raised to a large extent, and mouth corners obviously raised, the probability of corresponding to happy emotions is recorded as 0.8, the probability of corresponding to excited emotions is 0.2, and the probability of corresponding to other emotions is lower. Such mapping relationship data will serve as an important basis for subsequent emotional information reasoning and judgment (emotional states and certain health conditions affect each other and can serve as the basis for subsequent analysis).

[0034] Based on the facial feature data in the target face image information and the established knowledge base of the mapping relationship between facial features and emotions, the system uses reasoning algorithms (such as rule-based reasoning, Bayesian reasoning, etc.) to judge the user's emotional state. For example, when the target face image information shows that the user's eyes are moderately open, the eyebrows are slightly raised, and the corners of the mouth are raised to a certain extent, the system infers based on the mapping relationship knowledge base and judges that the user is in a positive emotional state (similar to mild happiness or satisfaction). Since the relationship between facial features and emotions is not absolutely one-to-one, there is a certain ambiguity and complexity. Therefore, the system will consider multiple characteristics when judging emotional states and give the confidence of each emotional state. For example, the system judges that the user has a probability of 0.6 in a positive emotional state (similar to happiness but not mentioning disease-related), a probability of 0.3 in a neutral emotional state (such as calm), and a probability of 0.1 in other unclassified emotional states. Such results can more comprehensively reflect the uncertainty of the user's emotional state (the accuracy of emotional judgment is of great significance for understanding the user's overall health and psychological state, and can be used as a basis for further evaluation).

[0035] If the system can obtain the user's continuous expression information over a period of time (for example, through the video stream captured by a continuous monitoring camera or multiple collections of facial images at different time points), the emotion judgment results at these different times can be fused and updated to obtain a more stable and accurate emotional state assessment. A common method is to use weighted average or dynamic Bayesian network and other technologies to dynamically adjust the judgment of the user's emotional state according to the trend and stability of expression changes in the time series. For example, if the user's expression characteristics gradually change from a slightly positive state (such as a slight upward turn of the mouth corners) to a more obvious positive state (such as wide eyes, raised eyebrows, and greatly raised mouth corners) over a period of time, the system will gradually increase the confidence that the user is in a positive emotional state (similar to happiness) based on this trend of change. At the same time, it can also combine other behavioral information of the user during this period (such as body language, voice intonation, etc., if relevant data is collected) to further enrich the source of emotional information and improve the accuracy of emotional judgment (multimodal information fusion is of great value for a comprehensive understanding of the user's state and can be used as a direction for subsequent expansion analysis).

[0036] The target user's identity attribute information is generated based on the target user's identity information, the target user's physiological type information and the target user's emotional information. The target user's identity attribute information includes the target user's identity information, the target user's physiological type information and the target user's emotional information. The target user's identity information (name, age, gender, contact information, etc.), physiological type information (preliminary judgment of minors, adults or elderly groups and gender information, etc.) and emotional information (the current emotional state judgment result obtained after the above processing) are integrated to form a complete target user's identity attribute information. In the hospital information system, ensure that the storage of this information in the database is consistent and relevant. Usually, a relational database structure is used, with the user's unique identifier (such as ID number or system-generated user ID) as the primary key, and the identity information, physiological type information and emotional information are stored in different fields or tables respectively, and the association relationship is established through the primary key, which is convenient for subsequent rapid query, call and analysis in various business modules (identity attribute information plays an important role in hospitals providing personalized services and optimizing resource allocation, and can be used as the basis for subsequent management decisions).

[0037] In terms of optimizing the customer service process in hospitals, personalized service guidance is provided based on the user's identity attribute information. For example, for the elderly (based on physiological type information), the system can automatically recommend more convenient medical routes, such as routes close to elevators and reducing walking distances, and provide special rest seats; for users with positive emotions (based on emotional information), some value-added service recommendations can be provided, such as health lectures and preferential activity information in the hospital; for users of different identities (such as ordinary users, medical insurance users, VIP users, etc., based on identity information), corresponding service treatment is provided.

[0038] S103: Process the medical record information of the target user and the physiological type information of the target user to generate user rating information.

[0039] In one implementation, feature extraction is performed on the target user's medical record information to generate medical record feature information, and the medical record feature information is processed to generate a medical record classification label and medication type information, wherein the medical record classification label is used to characterize the processing priority of the target user. Assume that the target user is a user suffering from multiple chronic diseases, and his medical record information includes many years of medical records, examination reports, diagnosis results, etc. First, feature extraction is performed on these medical record information. For example, the type of disease the user suffers from (such as hypertension, diabetes, etc.), the course of the disease (how many years of illness), and the severity index of the disease (such as the blood sugar control of diabetic users, the blood pressure fluctuation range of hypertensive users, etc.) are extracted from the medical record, and these extracted information constitute the medical record feature information. Then, the medical record feature information is processed to generate a medical record classification label and medication type information. For example, if the user's diabetes condition is poorly controlled and accompanied by multiple complications, a high-risk medical record classification label will be marked, indicating that the user needs priority treatment and close attention during the treatment process. At the same time, according to the user's disease condition, the type of medication is determined, such as hypoglycemic drugs, antihypertensive drugs, etc.

[0040] The physiological type information of the target user is subjected to feature extraction processing to generate physiological characteristic information. For the physiological type information of the target user, it is assumed that it includes basic information such as age, gender, height, and weight. Feature extraction processing is performed on these physiological type information. For example, the physiological stage of the user can be inferred from the age and gender information (such as middle-aged women facing menopause-related health problems), and indicators such as body mass index (BMI) can be calculated from the height and weight, which constitute physiological characteristic information. The physiological characteristic information is then processed to generate critical values ​​for tests, adverse reaction information, and physical function signs information. For example, for an obese (high BMI) user, it is more likely to have critical values ​​for dyslipidemia in blood tests, there is a higher risk of adverse reactions when taking certain drugs (such as effects on heart function), and physical function signs show decreased exercise capacity, shortness of breath, etc.

[0041] Physiological characteristic information is processed to generate critical values ​​for tests, adverse reaction information, and physical function signs information; medical record classification labels, medication type information, critical values ​​for tests, adverse reaction information, and physical function signs information are processed to generate user rating information. For example, a user with multiple chronic diseases, a high risk of critical values ​​for tests, adverse drug reactions, and poor physical function will be rated as a high-risk level and will require more frequent and detailed attention in subsequent medical monitoring; while a user with a milder condition, normal test indicators, no obvious risk of adverse reactions, and good physical function will be rated as a low-risk level. For child users (0-14 or 0-18 years old), considering that their physiological development is not yet mature, drug metabolism and response to diseases are different from those of adults, especially infants and young children, special attention needs to be paid to dosage adjustment and disease progression monitoring, which affects the rating. The elderly (over 60 years old, especially those over 85 years old) have declining physical functions, often accompanied by multiple chronic diseases, changes in tolerance and responsiveness to drugs, and weaker ability to recover from diseases. Therefore, a higher rating is given to the monitoring needs of this group of people during the rating process.

[0042] In another embodiment, the method further includes a calculation formula for obtaining a verification critical value, the calculation formula being:

[0043] Among them, C low represents the critical value comprehensive index below the lower limit, n represents the number of test indicators involved in the calculation, i is an index variable, from 1 to, used to traverse all test indicators, w i is the weight of the test index, which reflects the relative importance of different test indicators in assessing critical values. i is the normal lower limit of the i-th test index, x i is the actual measured value of the i-th test indicator, C high Indicates a comprehensive index of critical values ​​above the upper limit, U i is the normal upper limit value of the ith test index.

[0044] Assume that there are three test indicators involved in the calculation (i.e., n=3), namely blood glucose (x1), blood potassium (x2), and blood calcium (x3). Assume that the normal lower limit of blood glucose is L1=3.9mmol / L, the normal upper limit is U1=6.1mmol / L; the normal lower limit of blood potassium is L2=3.5mmol / L, the normal upper limit is U2=5.5mmol / L; the normal lower limit of blood calcium is L3=2.1mmol / L, the normal upper limit is U3=2.6mmol / L.

[0045] Assume that the weights of these three test indicators are w1=0.4, w2=0.3, and w3=0.3 respectively.

[0046] For the critical value comprehensive index C below the lower limit low :

[0047] Assume that the user's blood sugar measurement value x1=3.0mmol / L, blood potassium measurement value x2=3.2mmol / L, and blood calcium measurement value x3=2.0mmol / L.

[0048] According to the formula

[0049] For blood glucose: w1×max(0,L1-x1)=0.4×max(0,3.9-3.0)=0.4×0.9=0.36.

[0050] For blood potassium: w2×max(0,L2-x2)=0.3×max(0,3.5-3.2)=0.3×0.3=0.09.

[0051] For blood calcium: w3×max(0,L3-x3)=0.3×max(0,2.1-2.0)=0.3×0.1=0.03.

[0052] Then C low =0.36+0.09+0.03=0.48.

[0053] For the critical value comprehensive index C above the upper limit high :

[0054] Assume that the user's blood sugar measurement value x1=7.0mmol / L, blood potassium measurement value x2=6.0mmol / L, and blood calcium measurement value x3=2.8mmol / L.

[0055] According to the formula:

[0056] For blood glucose: w1×max(0,x1-U1)=0.4×max(0,7.0-6.1)=0.4×0.9=0.36.

[0057] For blood potassium: w2×max(0,x2-U2)=0.3×max(0,6.0-5.5)=0.3×0.5=0.15.

[0058] For blood calcium: w3×max(0,x3-U3)=0.3×max(0,2.8-2.6)=0.3×0.2=0.06.

[0059] Then C high =0.36+0.15+0.06=0.57.

[0060] These critical value indicators can help medical personnel determine whether the user's health condition is in a dangerous range that requires emergency treatment, so that they can take appropriate medical measures.

[0061] S104, preprocessing the training sample set to generate a training sample set with target feature data.

[0062] In one implementation, the training data set is grouped to generate a grouped training data set, wherein the grouped training data set includes physiological characteristic information of users of different age groups. Assume that we have a training data set containing physiological characteristic information of 1,000 users. The age range of these users ranges from 10 to 80 years old. We group the data set according to age groups, such as children's group (10-17 years old), youth group (18-35 years old), middle-aged group (36-59 years old) and elderly group (60 years old and above). In this way, a grouped training data set is obtained, and each group contains physiological characteristic information such as height, weight, heart rate, blood pressure, and blood routine of users in the corresponding age group. Feature extraction is performed on the grouped training data set to generate an original feature library. For each grouped data set, we perform feature extraction. Taking the youth group as an example, we will calculate statistical features such as the average height, average weight, standard deviation of heart rate, and average blood pressure of users in the group, as well as some derived features, such as the ratio of height to weight. All these features extracted from different groups are aggregated together to form an original feature library.

[0063] Divide the original feature library into various feature data sets to generate training sets and validation sets. Assume that there are 100 features in the original feature library. We divide the original feature library into training sets and validation sets according to a certain ratio (such as 70:30). For example, randomly select 70 features as the training set and the remaining 30 features as the validation set. The data in the training set will be used to train the model, while the validation set will be used to evaluate the performance of the model. Based on the classifier, predict the validation sets divided by the original feature library to generate prediction results. Use a classifier (such as a support vector machine, decision tree, etc.) to predict the validation set. Assuming that we are using a decision tree classifier, input the feature data in the validation set into the decision tree model. The model will classify and predict each sample based on the rules and patterns it has learned, such as predicting whether the user has physiological abnormalities. These prediction results are based on the validation set. Use the preset algorithm to divide the training sets in the original feature library for training and generate validation set class prediction results. Use the preset algorithm (such as the gradient descent algorithm for training neural networks, etc.) to train on the training set. During the training process, the model continuously adjusts its parameters to minimize the loss function. After the training is completed, the validation set is input into the trained model to obtain another set of prediction results, namely the validation set class prediction results.

[0064] The prediction results and the validation set prediction results are processed to generate target feature data, which is used to characterize the feature information that affects the user's physiological abnormalities. Compare the above two prediction results (the results of direct prediction of the validation set and the results of prediction of the validation set after training based on the training set). By analyzing the differences between the two results, such as observing which features play a key role in the two predictions and which features lead to inconsistencies in the prediction results, these key features are extracted as target feature data. These target feature data can characterize the feature information that affects the user's physiological abnormalities. For example, if it is found that the heart rate variability in the youth group is strongly correlated with physiological abnormalities, then the feature of heart rate variability will be included in the target feature data.

[0065] S105: Process the preset user information screening model based on the training sample set with target feature data to generate a target user information screening model.

[0066] In one implementation, suppose we are building a user information screening model for a hospital, with the goal of screening out users who need special medical attention. First, collect a large amount of user data from the hospital's electronic medical record system, physical examination center, and various medical testing equipment. These data include the user's basic information (age, gender, height, weight, etc.), medical history, family history, recent examination results (such as blood routine, biochemical indicators, imaging examination results), lifestyle habits (whether smoking, drinking frequency, amount of exercise, etc.), etc. For example, we collected data from 10,000 users as the initial training sample set. Through research and analysis by professional doctors and data analysts, determine the target feature data related to whether the user needs special medical attention. These features include: age over 60 and suffering from chronic diseases (such as hypertension, diabetes).

[0067] Some key indicators in recent examinations are abnormal, such as serum creatinine values ​​higher than the normal range (indicating abnormal renal function). There is a specific family history and related symptoms, such as multiple people in the family have heart disease, and the user has recently experienced symptoms such as palpitations and chest tightness. Extract samples containing these target feature data from the initial training sample set to form a training sample set with target feature data. Assume that after screening, 5,000 samples meet the requirements, and these samples constitute the data set for training the model.

[0068] Select a suitable machine learning model as the preset user information screening model, such as a logistic regression model. Logistic regression models are suitable for binary classification problems. In this scenario, they can be used to determine whether a user needs special medical attention (yes or no). Initialize the logistic regression model, including setting the parameters of the model (such as initializing the weight coefficient to a random value). Preprocess the data in the training sample set with the target feature data. This includes data normalization (mapping numerical features of different ranges to the same interval, such as normalizing all numerical features to the interval [0,1]), data encoding (such as numerically encoding categorical variables such as gender, with male as 0 and female as 1), etc., to ensure that the data is suitable for model training. Input the preprocessed training sample set into the preset user information screening model for training. In the logistic regression model, the training process is to minimize the loss function (such as the logarithmic loss function) through an optimization algorithm (such as the gradient descent method). Specifically, for each training sample, the model calculates the predicted probability that the user needs special medical attention based on the current parameters, and then adjusts the model parameters based on the difference between the predicted probability and the actual label (whether special medical attention is needed). This process is repeated until the model converges, that is, the loss function no longer decreases significantly.

[0069] Use a part of the training sample set with target feature data as the validation set (for example, take 1000 samples out of 5000 as the validation set) to evaluate the model in the training process. Evaluation indicators can be accuracy, recall, F1-score, etc. Optimize the model according to the evaluation results. For example, if it is found that the accuracy of the model on the validation set is low, it is necessary to adjust the parameters of the model, increase the amount of training data, or try other optimization algorithms. After multiple training and optimization, when the performance of the model on the validation set reaches a satisfactory level, we get the target user information screening model. This model can be used to screen new user data to determine whether they need special medical attention. For example, when a new user comes to the hospital, their data is input into the target user information screening model, and the model can quickly give a preliminary judgment to help the hospital allocate medical resources more efficiently and provide targeted medical services.

[0070] S106: Process the user rating information and the identity information of the target user based on the target user information screening model to generate initial monitoring information of the target user.

[0071] In one embodiment, the user rating information and the identity information of the target user are processed based on the target user information screening model to generate the real-time monitoring level information of the target user. Assume that the target user information screening model is a model built based on a machine learning algorithm (such as a decision tree). The model uses a large amount of user data in the training phase, including user rating information (such as disease severity, medication complexity, etc.) and identity information (such as age, gender, etc.) as input features, and outputs different monitoring levels (such as high, medium, and low). For example, the target user is a 70-year-old male who suffers from a variety of chronic diseases (such as hypertension, diabetes), and recent examinations show that some indicators are poorly controlled (this is reflected in the user rating information). These user rating information (high disease complexity, poor indicator control) and identity information (elderly male) are input into the target user information screening model. After internal calculation and judgment, the model outputs the real-time monitoring level information of the target user as "high" according to the rule: elderly people with multiple chronic diseases and poor recent indicators belong to high monitoring levels.

[0072] Get the real-time monitoring level information and real-time monitoring personnel information of other users. Assume that there are multiple users (users) in a hospital ward management system. Through the system, the real-time monitoring level information of other users can be obtained, such as the real-time monitoring level of user A is "medium" and the real-time monitoring level of user B is "low". At the same time, the system also records the real-time monitoring personnel information, such as user A is monitored by nurse A, and user B is monitored by nurse B. This information includes the professional skills and work schedule of the monitoring personnel. The real-time monitoring level information of the target user and the real-time monitoring level information of other users are processed to generate the monitoring priority of the target user. The real-time monitoring level of the target user ("high") is compared with the real-time monitoring levels of other users (such as "medium" for user A and "low" for user B). According to the pre-set rules, the higher the monitoring level, the higher the monitoring priority. So in this example, the monitoring priority of the target user is higher than that of user A and user B.

[0073] The real-time monitoring personnel information and the monitoring priority of the target user are processed to generate the initial monitoring information of the target user. Given the real-time monitoring personnel information, it is assumed that the hospital currently has nurses A, B, and C available for assignment. Nurse A is currently responsible for user A (medium monitoring level), nurse B is responsible for user B (low monitoring level), and nurse C is currently idle. According to the high monitoring priority of the target user, the system will consider assigning nurse C to the target user for monitoring. The initial monitoring information of the target user includes: the monitoring level is "high", the monitoring personnel is nurse C, and the monitoring tasks include close monitoring of the target user's chronic diseases (such as measuring blood pressure and blood sugar every hour), regularly evaluating the effect of medication, and arranging doctors to conduct more frequent ward rounds. These initial monitoring information will serve as the basis for subsequent detailed monitoring plans to ensure that the target user can receive timely and appropriate medical monitoring.

[0074] S107, processing the initial monitoring information of the target user based on the consultation purpose information of the target user and the emotional information of the target user to generate target monitoring information of the target user.

[0075] In one implementation, when the target user asks the hospital consultation platform, "I just had a heart bypass surgery and want to know how to do rehabilitation exercises after surgery?", the system does not simply identify keywords. It uses natural language processing technology to deeply analyze the grammatical structure and semantic logic of the entire sentence. For example, it understands that "just had surgery" means that it is in the early postoperative stage, and "rehabilitation exercises" is the focus. At the same time, the user's word usage habits, language style, etc. are analyzed to more comprehensively grasp the consultation intention. In addition to identifying direct keywords such as "heart bypass surgery" and "postoperative rehabilitation exercises", the system also associates related medical concepts. For example, heart bypass surgery involves the reconstruction of heart vessels, so related indirect concepts such as cardiovascular system recovery and myocardial function training will also be taken into consideration. Based on these comprehensive analyses, it is determined that the consultation is closely related to postoperative rehabilitation of the heart, and then it is classified as "postoperative rehabilitation monitoring of the heart". The preset classification rules are not static, but are continuously optimized based on the update of medical knowledge and clinical practice experience. For example, for postoperative rehabilitation monitoring of heart surgery, rehabilitation consultations for different surgical methods (such as arterial bypass and venous bypass) and different disease severity (such as single-vessel disease and multi-vessel disease) are further subdivided to provide more accurate monitoring classification. When encountering complex or ambiguous consulting content, the system will refer to a large number of historical consulting case libraries to find classification results of similar cases, and combine the opinions of medical experts to ensure the accuracy of the classification. For example, if the user asks "I feel a little short of breath after heart bypass surgery. Is it normal and how to improve it?", the system will look for similar postoperative symptom consultation cases in the case library and comprehensively judge its monitoring classification.

[0076] In order to more accurately judge the user's emotional state, the system does not rely solely on voice intonation and text expression. It also combines the user's facial expressions during the consultation (if there is a video consultation), the speed of inputting text, and pause time and other multimodal information for comprehensive analysis. For example, when the user is in a voice consultation, the voice trembles, the speed of speaking increases, and there are frequent pauses. At the same time, the repeated use of emphatic words in the text expression may all be manifestations of anxiety. Using emotional computing technology, the user's voice and text are scored for emotional tendencies. For example, emotions are divided into three dimensions: positive, negative, and neutral. Anxiety will have a higher score in the negative dimension. At the same time, the intensity of emotions, such as mild anxiety, moderate anxiety, and severe anxiety, will also be analyzed in order to more accurately formulate monitoring and early warning strategies.

[0077] Different users may have different reactions and behaviors to anxiety. The system will evaluate the specific impact that anxiety may have on the user based on the user's personal characteristics, such as age, gender, and personality traits (if relevant information is available). For example, for young and introverted users, anxiety may cause them to be more inclined to self-isolate and not actively cooperate with the rehabilitation plan; while for older and cheerful users, anxiety may manifest as over-reliance on medical staff and frequent questions. Based on these personalized assessments, more targeted monitoring warning information is generated. Take into account the long-term effects that anxiety may cause, such as the impact on sleep quality and appetite, which in turn affects physical recovery. The system will continue to track the user's emotional changes and their impact on the physical condition, and adjust the monitoring warning level and intervention measures in a timely manner. For example, if it is found that the user's anxiety continues to lead to lack of sleep, affecting the heart function recovery indicators, the monitoring warning level will be increased and the monitoring frequency will be increased.

[0078] When generating dynamic monitoring adjustment information, the system will comprehensively consider multiple factors in monitoring classification information and monitoring warning information. In addition to increasing the frequency of cardiac function monitoring, arranging psychologists for evaluation and counseling, and strengthening rehabilitation exercise guidance and supervision, personalized adjustment strategies will be formulated according to the specific situation of the user, such as family support and economic conditions. For example, if the user's family support is good, family members can be advised to participate in the rehabilitation process, such as accompanying exercise and supervising medication; if the economic situation permits, some auxiliary rehabilitation equipment or nutritional supplements can be recommended. Combined with the resource allocation of the hospital, reasonable arrangements for monitoring adjustment measures are made. For example, if the hospital's psychologist resources are limited, priority will be given to users with more severe anxiety and greater impact on rehabilitation; if rehabilitation equipment is tight, the use time will be reasonably allocated according to the user's condition and rehabilitation stage. At the same time, the professional expertise of different medical staff will be considered to match the most suitable rehabilitation doctor for guidance for the user. Dynamic monitoring adjustment information is not determined at one time, but is continuously updated according to the user's real-time situation. For example, if the user has new problems or symptoms during rehabilitation exercise, the system will immediately adjust the monitoring plan. At the same time, when performing monitoring tasks, medical staff will provide real-time feedback to the system on the observed user conditions, such as the user's acceptance of psychological counseling, the progress of rehabilitation exercises, etc. The system will further optimize and adjust the strategy based on these feedbacks. Establish an information interaction mechanism with other medical systems to obtain more information about users in order to adjust the monitoring plan more accurately. For example, connect with the hospital's nutrition department system to adjust the nutritional support plan in the rehabilitation plan according to the user's diet; cooperate with the community medical system to understand the user's living environment and community rehabilitation resources after discharge, and include them in the consideration of dynamic adjustment of monitoring.

[0079] Adjusting the frequency of blood pressure and heart rate monitoring is not just a simple increase in the number of times, but also the optimization of monitoring time points. For example, in addition to the regular morning and evening measurements, measurements before and after exercise at noon or in the afternoon are added to more comprehensively understand the user's cardiovascular function in different states. At the same time, according to the user's daily activity pattern, the measurement reminder time is personalized to ensure that the user can measure on time. In terms of electrocardiogram examination, in addition to adjusting the examination frequency, a more appropriate examination type will be selected according to the user's condition changes. For example, if the user is found to have signs of arrhythmia during the rehabilitation process, the conventional electrocardiogram examination will be upgraded to dynamic electrocardiogram monitoring to capture abnormal electrocardiogram signals in time. For rehabilitation exercise guidance, professional rehabilitation doctors will develop personalized exercise plans based on the user's physical recovery, including adjustments to exercise intensity, exercise methods and exercise time. For example, low-intensity walking and simple limb stretching exercises are mainly used in the early stage, and the exercise intensity is gradually increased as the body recovers, such as light aerobic exercise.

[0080] During the execution of the target monitoring information, the system will track and record the user's changes in various indicators, rehabilitation progress, emotional state and other information throughout the process. For example, a detailed electronic health record is established to record each blood pressure and heart rate measurement results, electrocardiogram examination report, completion of rehabilitation exercises and psychological assessment results. Based on these tracking data, the effectiveness of the monitoring plan is continuously evaluated, problems are discovered and optimized in a timely manner. The target monitoring information is reviewed and summarized regularly, and the long-term monitoring strategy is adjusted according to the user's overall rehabilitation situation. For example, if the user's rehabilitation progress is smooth over a period of time and various indicators are stable, the system will appropriately reduce the monitoring level and reduce unnecessary monitoring and intervention; if the user has repeated or new problems, it will re-evaluate and formulate a stricter monitoring plan to ensure that the user can get the most appropriate medical monitoring service and achieve the best rehabilitation effect. At the same time, the user's monitoring experience and data are sorted and analyzed to provide reference and reference for the subsequent monitoring of similar users, and continuously improve the entire monitoring management system.

[0081] In one embodiment, if Figure 2 As shown, the present application also provides a management device for automatically monitoring a user, including:

[0082] An acquisition module 201 is used to acquire the target user's biometric information, the target user's medical history information, the target user's consultation purpose information, a preset user information screening model and a training sample set;

[0083] The processing module 202 is used to process the biometric information of the target user to generate the identity attribute information of the target user, wherein the identity attribute information of the target user includes the identity information of the target user, the physiological type information of the target user and the emotional information of the target user; process the medical record information of the target user and the physiological type information of the target user to generate user rating information; pre-process the training sample set to generate a training sample set with target feature data, wherein the target feature data is used to characterize the risk factors affecting the physiological characteristics of the user; based on the training sample set with the target feature data, process the preset user information screening model to generate a target user information screening model; based on the target user information screening model, process the user rating information and the identity information of the target user to generate the initial monitoring information of the target user; based on the consulting purpose information of the target user and the emotional information of the target user, process the initial monitoring information of the target user to generate the target monitoring information of the target user.

[0084] 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 management method, electronic device, electronic device, and readable storage medium embodiment for automatically monitoring users, since they are basically similar to the above-mentioned management method embodiment for automatically monitoring users, the description is relatively simple, and the relevant parts can be referred to the partial description of the above-mentioned management method embodiment for automatically monitoring users.

Claims

1. A management method for automatically monitoring a user, characterized in that: include: Obtaining biometric information of the target user, medical record information of the target user, consultation purpose information of the target user, a preset user information screening model and a training sample set; Processing the biometric information of the target user to generate identity attribute information of the target user, wherein the identity attribute information of the target user includes identity information of the target user, physiological type information of the target user, and emotional information of the target user; Processing the medical record information of the target user and the physiological type information of the target user to generate user rating information; Preprocessing the training sample set to generate a training sample set with target feature data; Processing the preset user information screening model based on the training sample set with target feature data to generate a target user information screening model; Processing the user rating information and the identity information of the target user based on the target user information screening model to generate initial monitoring information of the target user; The initial monitoring information of the target user is processed based on the consultation purpose information of the target user and the emotional information of the target user to generate the target monitoring information of the target user.

2. The method according to claim 1, characterized in that The biometric information of the target user is processed to generate identity attribute information of the target user, including: Processing the biometric information of the target user to generate initial facial image information, identity information of the target user, and physiological type information of the target user; Processing the initial facial image information based on preset processing rules to generate expression feature influencing factors of the target user; Processing the influencing factors of the target user's expression characteristics to generate target face image information, wherein the target face image information includes expression change information of the target user; Processing the target face image information to generate emotion information of the target user; The identity attribute information of the target user is generated based on the identity information of the target user, the physiological type information of the target user and the emotional information of the target user.

3. The method according to claim 1, characterized in that Processing the medical record information of the target user and the physiological type information of the target user to generate user rating information includes: Performing feature extraction processing on the medical record information of the target user to generate medical record feature information; Processing the medical record characteristic information to generate a medical record classification label and medication type information, wherein the medical record classification label is used to represent the processing priority of the target user; Performing feature extraction processing on the target user's physiological type information to generate physiological feature information; Processing the physiological characteristic information to generate critical test values, adverse reaction information and body function sign information; The medical record classification label, the medication type information, the verification critical value, the adverse reaction information and the body function sign information are processed to generate user rating information.

4. The method according to claim 3, characterized in that Processing the medical record information of the target user and the physiological type information of the target user to generate user rating information also includes: The method also includes a calculation formula for obtaining a verification critical value, the calculation formula being: Among them, C low represents the critical value comprehensive index below the lower limit, n represents the number of test indicators involved in the calculation, i is an index variable, w i is the weight of the test indicator, L i is the normal lower limit of the i-th test index, x i is the actual measured value of the i-th test indicator, C high Indicates a comprehensive indicator of critical values ​​above the upper limit, U i is the normal upper limit value of the ith test index.

5. The method according to claim 1, characterized in that Preprocessing the training data set to generate a training data set with target feature data includes: Performing grouping processing on the training data set to generate a grouped training data set, wherein the grouped training data set includes physiological characteristic information of users of different age groups; Extract features from the grouped training data set to generate an original feature library; Dividing the original feature library into various feature data sets to generate a training set and a validation set; Based on the classifier, each validation set divided by the original feature library is predicted to generate a prediction result; Use the preset algorithm to divide each training set in the original feature library for training and generate the prediction results of the validation set class; The prediction result and the validation set prediction result are processed to generate target feature data, where the target feature data is used to characterize feature information that affects physiological abnormalities of the user.

6. The method according to claim 3, characterized in that The user rating information and the identity information of the target user are processed based on the target user information screening model to generate initial monitoring information of the target user, including: Processing the user rating information and the identity information of the target user based on the target user information screening model to generate real-time monitoring level information of the target user; Obtaining other users' real-time monitoring level information and real-time monitoring personnel information; Processing the real-time monitoring level information of the target user and the real-time monitoring level information of other users to generate a monitoring priority level for the target user; The real-time monitoring personnel information and the monitoring priority of the target user are processed to generate initial monitoring information of the target user.

7. The method according to claim 6, characterized in that The initial monitoring information of the target user is processed based on the consulting purpose information of the target user and the emotional information of the target user to generate target monitoring information of the target user, including: Processing the consulting purpose information of the target user to generate guardianship classification information; Processing the target user's emotional information to generate monitoring warning information; Processing the monitoring classification information and the monitoring early warning information to generate monitoring dynamic adjustment information; The initial monitoring information of the target user is processed based on the dynamic monitoring adjustment information to generate target monitoring information of the target user.

8. A management device for automatically monitoring a user, characterized in that: The device comprises: An acquisition module, used to acquire the target user's biometric information, the target user's medical record information, the target user's consultation purpose information, a preset user information screening model and a training sample set; A processing module is used to process the biometric information of the target user to generate the identity attribute information of the target user, wherein the identity attribute information of the target user includes the identity information of the target user, the physiological type information of the target user and the emotional information of the target user; process the medical record information of the target user and the physiological type information of the target user to generate user rating information; pre-process the training sample set to generate a training sample set with target feature data, wherein the target feature data is used to characterize the risk factors affecting the physiological characteristics of the user; based on the training sample set with target feature data, process the preset user information screening model to generate a target user information screening model; based on the target user information screening model, process the user rating information and the identity information of the target user to generate the initial monitoring information of the target user; based on the consulting purpose information of the target user and the emotional information of the target user, process the initial monitoring information of the target user to generate the target monitoring information of the target user.

9. An electronic device, characterized in that: include: a first processor; and a memory for storing executable instructions of the first processor; Wherein, the first processor is configured to execute the management method for automatically monitoring a user 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 management method for automatically monitoring a user as described in any one of claims 1 to 7 is implemented.

Images