Smart medical interaction method and system based on big data

Abstract

The invention relates to the technical field of data interaction, and particularly provides a smart medical interaction method and system based on big data. The method comprises the steps that a physiological knowledge data model is created, a user inputs personal attributes through a terminal for registration, and different permissions are given according to the personal attributes; a user inputs related health data through the terminal and collects physiological data of a patient; related health data are analyzed, and weight values are given to the real-time physiological parameters; carrying out calibration according to the weight value; inputting the real-time physiological parameters into an illness state simulation model for simulation; and inputting a simulation result into the illness state simulation physiological knowledge data model, searching a corresponding doctor for answering according to the simulation result, and storing an answering process. The system comprises a health data acquisition module, a health data calling module and a health data interaction module. According to the invention, the quality and efficiency of medical services are improved, and an accurate, convenient and personalized health management scheme is provided for patients.

Description

Technical Field

[0001] The present invention relates to the field of data interaction technology, and in particular to a smart medical interaction method and system based on big data. Background Art

[0002] Smart medical interaction refers to the use of modern information technology and artificial intelligence technology to build an information-based and intelligent medical environment to achieve interaction between patients and medical staff, medical institutions, and medical equipment, optimize the medical process, and improve the quality and efficiency of medical services. The design scheme of smart medical interaction usually includes intelligent appointment system, intelligent diagnosis system, intelligent triage system, intelligent follow-up system, voice interaction technology, mobile application and intelligent data analysis. In smart medical interaction, the application of digital human technology has significantly improved the convenience and intelligence level of hospital services. For example, Moko Intelligent AI Digital Human Interaction All-in-One Machine helps patients quickly find the hospital destination through voice interaction, and provides functions such as registration, doctor query and patient question answering, thereby improving the medical experience. In addition, Yunzhisheng Company has comprehensively upgraded the intelligence level of various products in the medical business line based on the Shanhai Big Model, realizing the transition from assistant to expert. Smart medical interaction also involves a variety of intelligent communication forms, such as intelligent consultation, intelligent medical treatment and intelligent follow-up, which improve the patient experience by simplifying the process, improving efficiency and strengthening doctor-patient communication. Video consultation in telemedicine is an important scenario to promote communication between doctors and patients. With the help of information technology such as 5G and artificial intelligence, patients can enjoy real-time consultation services at home. However, smart medical interaction also faces some challenges, such as data security and privacy protection. Therefore, relevant national departments have issued a series of policies to promote the development of smart medical care and promote the intercommunication and sharing of health information through continuous innovation. In the future, the development of smart medical care requires multi-party collaboration, including the participation of governments, hospitals, colleges and universities, equipment manufacturers and information technology companies, to carry out industry-university-research cooperation, establish unified standards, continuously integrate and utilize cutting-edge innovative technologies, and promote the digitalization of the entire medical industry. Through the application of modern information technology and artificial intelligence technology, smart medical interaction not only improves the quality and efficiency of medical services, but also improves patients' medical experience, and promotes the doctor-patient relationship from the traditional "doctor-centered" communication model to the new "patient-centered" communication model. With the continuous advancement of science and technology, smart medical interaction will play a greater role in the future and provide patients with a better and more convenient medical service experience.

[0003] Prior art 1, Chinese patent, application number: 202410809501.8 discloses a method for health and medical data authentication and privacy protection in smart medical care, including the following steps: 1. The gateway device calls the key generation algorithm based on security parameters to output a public key for signature verification and a private key for signature; 2. The gateway device calls the signature algorithm based on the private key generated by the key generation algorithm and the health and medical data collected by the gateway device to output the signature of the health and medical data. Although the method for health and medical data authentication and privacy protection in smart medical care not only improves the flexibility of data sharing, but also prevents excessive editing of data, enhances the security and availability of shared health and medical data, and can provide more efficient security protection for the sharing of health and medical data in smart medical systems; however, the real-time response and processing capabilities are insufficient, resulting in the inability to process in a timely manner.

[0004] Prior art 2, Chinese patent, application number: 202410168178.0 discloses an Internet big data smart medical system based on remote wireless communication, which relates to the field of wireless communication technology. When the system is running, the logged-in user is verified through the client module, and the relevant information is filled in through the discomfort information module, including filling items and selection items, wherein the sequence information of the selection items is used as the first data set, and the user-filled information is processed by the data processing module, and the relevant discomfort characteristic data is counted as the second data set. By comparing with a large amount of patient health data stored in the cloud service storage module, the matching index Ppzs is obtained, and the matching index Ppzs and the first data set, as well as the second data set, are calculated by the information calculation module to obtain: the recommended index Jyzs, and the level recommended plan is obtained by comparing with the preset threshold, and the recommended document guide is formed according to the content of the plan. Although a clear action plan is provided for patients so that they can better manage their health; however, data sharing and integration are difficult, resulting in low service efficiency.

[0005] Prior art three, Chinese patent, application number: 202410149558.X discloses a smart medical system based on artificial intelligence, and the operation method of the system includes the following steps: Step 1: realize data collection and integration of medical equipment through sensors; Step 2: establish an abnormality detection model and realize fault prediction; Step 3: user interface and remote support; Step 4: predictive resource management, the data collection and integration module is used to collect medical equipment data and integrate data format; the abnormality detection and fault prediction module is used to detect abnormalities of medical equipment and realize abnormal fault prediction; the maintenance suggestion and optimization strategy module is used for the system to give maintenance suggestions and optimization plans for equipment, the abnormality detection module is used to timely discover abnormal conditions of equipment; the data integration and cleaning module is used to ensure the consistency and comparability of data. Although it has the characteristics of accurate prediction of medical equipment faults and effective improvement of maintenance efficiency; however, the real-time response and processing capabilities are insufficient, resulting in the inability to handle them in a timely manner.

[0006] At present, the existing technologies 1, 2 and 3 have insufficient real-time response and processing capabilities and difficulties in data sharing and integration. Therefore, the present invention provides a smart medical interaction method and system based on big data. Summary of the invention

[0007] In order to achieve the above object, the present invention adopts the following technical scheme:

[0008] In one aspect of the present invention, a smart medical interaction method based on big data is provided, comprising the following steps:

[0009] A physiological knowledge data model is created based on the historical physiological knowledge database. Users enter personal attributes through the terminal to register and are granted different permissions based on their personal attributes. Users enter relevant health data through the terminal to collect patient physiological data, and the collected patient physiological data is transferred to the physiological knowledge data model.

[0010] Analyze relevant health data and assign weight values ​​to real-time physiological parameters based on the analysis results; calibrate the physiological knowledge data model based on the weight values, and directly retrieve it if the calibration is successful; if the calibration fails, a personal data set is formed; divide the personal data set into a training set, a validation set, and a test set;

[0011] A disease state simulation model is constructed based on the patient's physiological database, and the disease state simulation model is trained using a training set; real-time physiological parameters are input into the disease state simulation model for simulation; the simulation results are input into the disease state simulation physiological knowledge data model, and the corresponding physician is searched for answers based on the simulation results, and the answer process is stored.

[0012] In an optional implementation, the process of collecting patient physiological data includes the following steps:

[0013] Create a physiological knowledge data model based on the historical physiological knowledge database, identify and connect various health data sources related to the research; obtain historical data from the selected data source; integrate different historical data into different physiological knowledge sub-datasets, and merge different physiological knowledge sub-datasets into a physiological knowledge dataset;

[0014] The user registers by inputting personal attributes through the terminal, and is granted different permissions according to the personal attributes; the user inputs relevant health data through the terminal, collects the patient's physiological data, and transmits the collected patient's physiological data to the physiological knowledge data model;

[0015] The patient's physiological data is cleaned, converted and formatted, and missing values ​​and outliers are processed, etc.; the preprocessed patient physiological data is input into the physiological knowledge data model for screening.

[0016] In an optional implementation, the process of calibrating the physiological knowledge data model according to the weight value comprises the following steps:

[0017] Compare and analyze relevant health data with historical user data to obtain the patient's health status; assign weight values ​​based on the patient's health status, and retrieve the data retrieval strategy and target retrieval database of the physiological knowledge data model according to the weight values;

[0018] Determine a database strategy of a target access database; retrieve target access data from the target access database based on the data access strategy and the database strategy; output the target access data to a user terminal;

[0019] If the retrieval fails, the calibration is judged to have failed; the relevant health data is integrated with the historical relevant data and the comparative analysis results to form a personal data set.

[0020] In an optional implementation, the process of obtaining the patient's health condition comprises the following steps:

[0021] Retrieve the user's historical status data, perform difference calculation with relevant health data, and obtain the status change value. If the status change value is positive, it is judged to be in a recovery state; if the status change value is negative, it is judged to be in a deterioration state; if the status change value is zero, it is judged to be in a stable state;

[0022] The state change value and the personal attribute value are summed up to calculate the patient's health status; the patient's health status is analyzed to obtain the patient's risk value; and the relevant health data are assigned a weight value according to the patient's risk value;

[0023] The data retrieval strategy and target retrieval strategy of the physiological knowledge model are retrieved based on the weight value; the data retrieval strategy is determined to be a paged retrieval strategy or a non-paged retrieval strategy, and the corresponding target application is determined.

[0024] In an optional implementation manner, the process of retrieving target retrieval data in the target retrieval database includes the following steps:

[0025] If it is determined that the data retrieval strategy is a paging retrieval strategy, then determine whether there is an address selection statement in the first call statement; if it is determined that there is a restriction condition statement in the first call statement, then determine whether the number of data return rows in the restriction condition statement is greater than a preset threshold;

[0026] If it is greater than the preset threshold, the target application is determined according to the application information, and the target retrieval data of the target application is retrieved from the target retrieval database according to the target application and the database policy; if the data retrieval policy is determined to be a non-paged retrieval policy, the retrieval data of the target application is retrieved from the target retrieval database according to the data query volume of the target application;

[0027] If it is determined that there is no restrictive condition language in the first calling statement, a preset restrictive statement is added to the first calling statement to obtain an updated second calling statement; a target database of the target application is called in the target calling database, and the called data is input into the user terminal.

[0028] In an optional implementation manner, the process of forming a personal data set includes the following steps:

[0029] Determine whether the preset application whitelist in the target database can match the target application; if it is determined that the target application can be matched in the preset application whitelist, determine the maximum number of return rows of the preset application whitelist; retrieve the target retrieved data of the target application in the target retrieved database according to the maximum number of return rows;

[0030] If the retrieval fails, it is judged as a calibration failure; the relevant health data is integrated with the historical relevant data and the comparative analysis results to form a personal data set;

[0031] The personal data set is preprocessed by removing outliers, balancing data distribution, converting categorical variables, and standardizing numerical variables; the personal data set is divided into training set, validation set, and test set.

[0032] In an optional implementation, the process of inputting the simulation results into the disease state simulation physiological knowledge data model comprises the following steps:

[0033] Retrieving historical disease data from the patient's physiological database based on the personal data set; using the training set in the personal data set to train the disease state model; during the training process, the personal data set is analyzed with the user's historical disease data to obtain the disease data trend;

[0034] The disease data is actually learned, and variables are screened recursively. Based on the training results of each disease state model, variables with values ​​less than the preset value are gradually eliminated. According to the simulation results of the final training disease state model, the variables that contribute the most to the disease state simulation results are extracted.

[0035] The real-time physiological parameters are input into the disease state simulation model for simulation; the simulation results are input into the physiological knowledge data model; the physiological knowledge data model outputs the simulation results to the user terminal; and the corresponding physician is searched for answers based on the simulation results, and the answers are stored.

[0036] In an optional embodiment, the process of extracting the variable that contributes most to the disease state simulation result comprises the following steps:

[0037] Collect the node set of the training set and the user's historical disease data in the personal data set, analyze the output data and output data characteristics of each node in the node set, and analyze the dependency relationship between nodes based on the characteristic analysis results of the node input data and output data;

[0038] According to the dependency relationship, a dependency matrix between nodes is constructed, and the similarity between the output data of the historical diseased node and the output data of the real-time healthy data node is calculated. According to the threshold between the similarities, the dependency relationship between the output data of the historical diseased node and the data of the real-time healthy data node is determined;

[0039] According to the dependency relationship, the node configuration information that can be optimized, as well as the potential parallel execution sets and sequential execution sets between the nodes are identified; according to the identified parallel execution sets and sequential execution sets, the execution order of the dependency configuration is reordered and optimized.

[0040] In an optional implementation manner, the process of outputting the simulation results to the user terminal includes the following steps:

[0041] Based on each training result of the diseased state simulation model, the variables smaller than the threshold are gradually eliminated, and according to the final training diseased state simulation result, the variable value that contributes the most to the simulation result is extracted; and the final diseased state simulation model is obtained;

[0042] Input the real-time physiological parameters into the disease state simulation model, simulate the real-time physiological parameters, and obtain the user's health data changes; compare the user's health data changes with the preset health data changes;

[0043] The comparison results are evaluated and input into the physiological knowledge data model; the evaluation results are output to the user terminal, and the corresponding physician is searched for answers based on the evaluation results, and the answering process is stored.

[0044] Another aspect of the present invention provides a smart medical interaction system based on big data, comprising:

[0045] The health data collection module is used to create a physiological knowledge data model based on the historical physiological knowledge database. Users enter personal attributes through the terminal to register and are granted different permissions based on their personal attributes. Users enter relevant health data through the terminal to collect patient physiological data, and the collected patient physiological data is transferred to the physiological knowledge data model.

[0046] Retrieve the health data module to analyze the relevant health data and assign weight values ​​to the real-time physiological parameters according to the analysis results; calibrate the physiological knowledge data model according to the weight values, and directly retrieve it if the calibration is successful; if the calibration fails, form a personal data set; divide the personal data set into a training set, a validation set, and a test set;

[0047] The interactive health data module is used to build a disease state simulation model based on the patient's physiological database, and use the training set to train the disease state simulation model; input real-time physiological parameters into the disease state simulation model for simulation; input the simulation results into the disease state simulation physiological knowledge data model, find the corresponding doctor for answer according to the simulation results, and store the answer process.

[0048] Through user registration and input of personal attributes, the system of the present invention can provide personalized health management plans for different users; the collected health data can update the physiological knowledge data model in real time to improve the accuracy and timeliness of health monitoring. Through real-time data analysis and weight assignment, the system can dynamically adjust the physiological knowledge data model to improve the adaptability and accuracy of the model; dividing the personal data set into training set, verification set and test set helps to improve the training effect and generalization ability of the model. Through the disease state simulation model, the system can more accurately predict the patient's disease state and improve the accuracy of diagnosis; the system can automatically find the corresponding physician for answer based on the simulation results and provide intelligent medical services; the answer process is stored to help accumulate knowledge and optimize future service processes. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0050] Figure 1A flowchart of the steps of providing a smart medical interaction method based on big data in Embodiment 1 of the present invention;

[0051] Figure 2 A flowchart of the steps of collecting patient physiological data in a smart medical interaction method based on big data provided in Example 2 of the present invention;

[0052] Figure 3 A flow chart of the steps of calibrating the physiological knowledge data model according to the weight value of the smart medical interaction method based on big data provided in Example 3 of the present invention;

[0053] Figure 4 A flowchart of the steps of obtaining the patient's health status in the smart medical interaction method based on big data provided in Example 4 of the present invention;

[0054] Figure 5 A flowchart of the steps of retrieving target data in a target retrieval database of a smart medical interaction method based on big data provided in Example 5 of the present invention;

[0055] Figure 6 A flowchart of the steps of forming a personal data set in a smart medical interaction method based on big data provided in Example 6 of the present invention;

[0056] Figure 7 A flowchart of the steps of inputting simulation results into a diseased state simulation physiological knowledge data model for providing a smart medical interaction method based on big data in Example 7 of the present invention;

[0057] Figure 8 A flowchart of the steps of extracting the variable that contributes most to the disease state simulation result in the smart medical interaction method based on big data in Example 8 of the present invention is provided;

[0058] Fig. 9 A flowchart of the steps of outputting simulation results to a user terminal in the smart medical interaction method based on big data provided in Example 9 of the present invention;

[0059] Fig.10 A functional module diagram of a smart medical interaction system based on big data is provided in Embodiment 10 of the present invention;

[0060] Fig.11 is a block diagram of an electronic device provided in Embodiment 11 of the present invention;

[0061] Fig.12 This is a block diagram of the computer-readable storage medium provided in Example 12 of the present invention. DETAILED DESCRIPTION

[0062] The technical solutions in the embodiments of the present invention will be described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0063] In the following, the terms "first", "second", etc. are used only for convenience of description and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first", "second", etc. may explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise specified, "plurality" means two or more.

[0064] In the present invention, unless otherwise clearly specified and limited, the term "connection" should be understood in a broad sense, for example, "connection" can be a fixed mechanical connection, or a detachable mechanical connection, or integrated; or, "connection" can be a direct connection, or an indirect connection through an intermediate medium. In addition, unless otherwise clearly specified and limited, the term "coupling" should be understood in a broad sense, for example, "coupling" can be a direct electrical connection, such as physical contact and electrical conduction between two components, and can also be understood as electrical connection between different components in a circuit structure through physical lines such as copper foil or wires on a printed circuit board (PCB) that can transmit electrical signals to transmit electrical signals; or, "coupling" can be an indirect electrical connection between two components through an intermediate medium; or, "coupling" can be an electrical connection between two components in an air-spaced / non-contact manner, for example, two components are electrically connected by capacitive coupling to transmit electrical signals.

[0065] In the embodiments of the present invention, directional terms such as "up", "down", "left" and "right" may be defined including but not limited to the orientation relative to the schematic placement of the components in the drawings. It should be understood that these directional terms may be relative concepts, which are used for relative description and clarification, and may change accordingly according to the change of the orientation of the components in the drawings.

[0066] Embodiments of the present invention can be used in smart hospitals and wards, telemedicine and remote monitoring, medical resource optimization and management, smart medical platforms and cloud services, and mobile medical and wearable devices.

[0067] Key technical features of the embodiments of the present invention are as follows: In the embodiments of the present invention, the system can assign different permissions to different users through the user inputting personal attributes and health data, and build a personalized physiological knowledge data model based on personal data; the system can collect the patient's physiological data (such as heart rate, blood pressure, blood sugar level, etc.) in real time, and perform real-time analysis based on these data. By assigning physiological parameter weight values ​​and calibrating the physiological knowledge data model, the system can dynamically adjust the model parameters; the disease state simulation model built based on the patient's physiological database can be learned through the training set, and simulate the real-time physiological parameters to predict the patient's disease state; the system supports different users to open different clinical data access rights; through role management and system log records, user operations can be effectively managed to prevent data leakage and abuse; the system can find the corresponding physician for answers based on the simulation results, and store the answer process; the system can integrate physiological data from different devices and sensors, analyze and mine through big data processing technology, discover the potential laws and trends behind the data, and provide users with valuable health information and suggestions; the system can predict future changes in health indicators based on real-time physiological indicator data, and generate corresponding dynamic tags. Combined with user portrait information, the system can recommend personalized health knowledge to help users better manage their health.

[0068] Embodiment 1:

[0069] like Figure 1 As shown, the embodiment of the present invention provides a smart medical interaction method based on big data, comprising the following steps:

[0070] Step S100: A physiological knowledge data model is created based on a historical physiological knowledge database. The user enters personal attributes through a terminal to register, and different permissions are granted according to the personal attributes. The user enters relevant health data through the terminal, collects the patient's physiological data, and transmits the collected patient's physiological data to the physiological knowledge data model.

[0071] Among them, different permissions include patient permissions, doctor permissions, and management permissions;

[0072] The expression of the physiological knowledge data model is:

[0073]

[0074] Activation function calculation formula:

[0075]

[0076] The calculation formula of authority coefficient is:

[0077]

[0078] The trend coefficient calculation formula is:

[0079]

[0080] Where n represents the number of data sources, that is, how many different health data sources are integrated into the model; D i′ represents the i′th data source, such as physiological parameters such as heart rate, blood pressure, blood sugar level, or primary care practice, laboratory data, insurance data, etc.; W i′ represents the importance or contribution of the i′th data source in the model. Different data sources may have different impacts on the model, so different weights need to be assigned. U represents the user's personal attributes, such as age, gender, medical history, etc. These attributes will affect the analysis results of the model. P represents the permissions granted to the user based on his or her attributes. The permission coefficient will affect the scope and depth of data that the user can access or use. m represents the number of time series, that is, how many time series data are used in the model. T j represents the jth time series data, such as the change of heart rate over time, the change of blood pressure over time, etc.; TC j represents the trend coefficient of the jth time series data, which is used to capture the changing trend of the time series data; p represents the number of neural network layers, that is, how many layers of neural networks are included in the model; L k Represents the k-th layer of the neural network, which is used to perform nonlinear transformation and feature extraction on the data; AF k represents the activation function of the k-th layer of the neural network, such as the Sigmoid function, which is used to introduce nonlinear characteristics; q represents the number of data fusions, that is, how many different data fusion methods are used in the model; F l Indicates the first data fusion method, such as weighted average, principal component analysis, etc., which is used to fuse data from different sources; FW l represents the weight of the lth data fusion method, which is used to adjust the contribution of different data fusion methods in the model; e represents a natural constant, which is approximately equal to 2.71828; r represents the total number of user attributes; A u represents the weight of the u-th user attribute, indicating the degree of influence of this attribute on the permission; C u represents the contribution value of the u-th user attribute, indicating the specific value or importance of the attribute; t represents the index of the time point, s represents the total number of time points of the time series data; T j G(t) represents the value of the jth time series at time point t, such as the value of heart rate, blood pressure, etc. that changes over time; G(t) represents the weight or importance of time point t, which is used to adjust the impact of different time points on the trend;

[0081] Step S200: Analyze the relevant health data, assign weight values ​​to the real-time physiological parameters according to the analysis results; calibrate the physiological knowledge data model according to the weight values, and directly retrieve it if the calibration is successful; if the calibration fails, form a personal data set; divide the personal data set into a training set, a validation set, and a test set;

[0082] Among them, physiological data includes heart rate, blood pressure, blood sugar level, and medication intake;

[0083] Step S300: construct a disease state simulation model based on the patient's physiological database, and use the training set to train the disease state simulation model; input the real-time physiological parameters into the disease state simulation model for simulation; input the simulation results into the disease state simulation physiological knowledge data model, find the corresponding doctor for answer according to the simulation results, and store the answer process.

[0084] In the above embodiment, in step S100 of this embodiment, a data model that can reflect the relationship between physiological parameters and health status is constructed based on the historical physiological knowledge database; the user enters personal attributes through the terminal to register, and the system grants different permissions (patient permissions, doctor permissions, management permissions, etc.) according to these attributes; the user enters relevant health data through the terminal, and these data are collected and transmitted to the physiological knowledge data model; in this embodiment, through user registration and input of personal attributes, the system can provide personalized health management plans for different users; the collected health data can update the physiological knowledge data model in real time to improve the accuracy and timeliness of health monitoring. Step S200 analyzes the collected health data and assigns weight values ​​to real-time physiological parameters according to the analysis results; the physiological knowledge data model is calibrated according to the weight values, and if the calibration is successful, it is directly retrieved; if the calibration fails, a personal data set is formed, and the personal data set is divided into a training set, a validation set and a test set; in this embodiment, through real-time data analysis and weight assignment, the system can dynamically adjust the physiological knowledge data model to improve the adaptability and accuracy of the model; dividing the personal data set into a training set, a validation set and a test set helps to improve the training effect and generalization ability of the model. In step S300, a disease state simulation model is constructed based on the patient's physiological database; the disease state simulation model is trained using a training set, and real-time physiological parameters are input into the disease state simulation model for simulation; the simulation results are input into the disease state simulation physiological knowledge data model, and the corresponding physician is searched for answers based on the simulation results, and the answer process is stored; in this embodiment, through the disease state simulation model, the system can more accurately predict the patient's disease state and improve the accuracy of diagnosis; the system can automatically search for the corresponding physician for answers based on the simulation results and provide intelligent medical services; storing the answer process helps to accumulate knowledge and optimize future service processes.

[0085] Embodiment 2:

[0086] like Figure 2 As shown, based on Example 1, the process of collecting patient physiological data in step S100 provided in this embodiment of the present invention includes the following steps:

[0087] Step S101: creating a physiological knowledge data model based on the historical physiological knowledge database, identifying and connecting various health data sources related to the research; obtaining historical data from the selected data source; integrating different historical data into different physiological knowledge sub-data sets, and merging different physiological knowledge sub-data sets into a physiological knowledge data set;

[0088] The various sources of health data include primary care practices, laboratory data, and insurance data; the selected data sources include physiological parameters such as heart rate, blood pressure, and blood sugar levels;

[0089] Step S102: The user enters personal attributes through the terminal to register, and different permissions are granted according to the personal attributes; the user enters relevant health data through the terminal, collects the patient's physiological data, and transmits the collected patient's physiological data to the physiological knowledge data model;

[0090] Step S103: pre-processing the patient's physiological data, such as cleaning, converting and formatting, and processing missing values ​​and abnormal values; inputting the pre-processed patient's physiological data into the physiological knowledge data model for screening.

[0091] In the above embodiment, in step S101 of this embodiment, based on the historical physiological knowledge database, a data model that can identify and connect various sources of health data is constructed; sources of health data including primary care practices, laboratory data, and insurance data are identified and connected; physiological parameters such as heart rate, blood pressure, and blood sugar levels are selected as data sources; historical data are obtained from the selected data source; different historical data are integrated into different physiological knowledge sub-data sets, and then these sub-data sets are merged into a complete physiological knowledge data set; this embodiment improves the integrity and availability of data by integrating data from different sources to form a comprehensive physiological knowledge data set; ensures that data from different sources have the same format and structure to facilitate subsequent data processing and analysis; the constructed data model can effectively identify and connect various sources of health data, providing a solid foundation for subsequent data analysis; through integration and standardization, the quality of data is improved, and the possibility of data inconsistency and errors is reduced. In step S103, the user registers by inputting personal attributes through the terminal, and different permissions are granted according to personal attributes; the user inputs relevant health data through the terminal to collect the patient's physiological data; the collected patient's physiological data is transmitted to the physiological knowledge data model; this embodiment ensures the security and privacy of data through user registration and authority management; realizes the real-time collection of patient's physiological data, improves the timeliness and accuracy of data; ensures that the data can be transmitted to the data model in time, and provides support for subsequent data processing and analysis. In step S103, the patient's physiological data is cleaned, duplicate data is removed, missing data is filled, and erroneous data is corrected; the data is converted, including data normalization, data segmentation, etc.; the data is formatted to ensure the consistency and readability of the data; missing values ​​and abnormal values ​​are processed to improve the quality of the data, and the pre-processed patient's physiological data is input into the physiological knowledge data model for screening. This embodiment improves the quality of data and reduces the impact of noise and errors through data cleaning and preprocessing; and ensures the consistency and readability of data through data conversion and formatting to facilitate subsequent data analysis. It improves the reliability and accuracy of data by processing missing values ​​and outliers. It ensures that only high-quality data is input into the data model through screening, thereby improving the performance and accuracy of the model.

[0092] Embodiment 3:

[0093] like Figure 3 As shown, based on Example 1, the process of calibrating the physiological knowledge data model according to the weight value in step S200 provided in the embodiment of the present invention includes the following steps:

[0094] Step S201: Compare and analyze the relevant health data with the user's historical data to obtain the patient's health status; assign a weight value based on the patient's health status, and retrieve the data retrieval strategy and target retrieval database of the physiological knowledge data model according to the weight value;

[0095] Step S202: determining a database strategy of a target retrieval database; retrieving target retrieval data from the target retrieval database based on the data retrieval strategy and the database strategy; and outputting the target retrieval data to a user terminal;

[0096] Step S203: If the retrieval fails, it is determined that the calibration has failed; the relevant health data is integrated with the historical relevant data and the comparative analysis results to form a personal data set; the personal data set is divided into a training set, a validation set and a test set.

[0097] In the above embodiment, in step S201 of this embodiment, by comparing and analyzing the current health data with the historical data, the patient's health change trends and potential problems can be identified. This step utilizes data mining and machine learning techniques, such as feature selection and feature extraction, to identify key features related to the patient's health status; according to the analysis results, a weight value is assigned to each feature or data point, which usually involves feature importance assessment and parameter adjustment during model training; based on the weight value, a suitable physiological knowledge data model and target database are selected to ensure that the retrieved data can effectively support subsequent health analysis and decision-making; this embodiment can more accurately identify the patient's health status and reduce the risk of misdiagnosis and missed diagnosis through comparative analysis and weight assignment; selecting a suitable data model and database according to the weight value can improve the efficiency and accuracy of data retrieval and avoid unnecessary waste of resources. In step S202, a suitable database access strategy is formulated according to the data retrieval strategy to ensure the integrity and consistency of the data; the data retrieval operation is performed in the target database, which usually involves SQL query, data interface call and other technologies; the retrieved data is output to the user terminal, which may involve data format conversion, data visualization and other technologies; this embodiment can quickly respond to user data needs and improve user experience by optimizing database strategies and retrieval processes; ensure that the data retrieved from the database is consistent with the data displayed on the user terminal to avoid information loss or errors. If data retrieval fails in step S203, calibration failure is determined, and corresponding error handling measures are taken; relevant health data are integrated with historical data to form a complete personal data set; relevant health data are integrated with historical data to form a complete personal data set; the preprocessed data set is divided into a training set, a validation set, and a test set to facilitate subsequent model training and evaluation; this embodiment improves the quality and reliability of data through data cleaning and preprocessing, provides a solid foundation for subsequent modeling and analysis, and through reasonable division of the data set, can effectively evaluate the performance of the model and prevent the occurrence of overfitting; by integrating and preprocessing personal health data, can better support the formulation and implementation of personalized health management plans.

[0098] Embodiment 4:

[0099] like Figure 4 As shown, based on Example 3, the process of obtaining the patient's health status in step S201 provided in this embodiment of the present invention includes the following steps:

[0100] Step S2011: retrieve the user's historical status data, perform difference calculation with the relevant health data, and obtain a status change value. If the status change value is positive, it is judged to be a recovery state; if the status change value is negative, it is judged to be a deterioration state; if the status change value is zero, it is judged to be in a stable state;

[0101] Step S2012: summing up the state change value and the personal attribute value to obtain the patient's health status; analyzing the patient's health status to obtain the patient's risk value; assigning a weight value to the relevant health data according to the patient's risk value;

[0102] Step S2013: retrieve the data retrieval strategy and target retrieval strategy of the physiological knowledge model based on the weight value; determine whether the data retrieval strategy is a paged retrieval strategy or a non-paged retrieval strategy, and determine the corresponding target application.

[0103] In the above embodiment, in step S2011 of this embodiment, the historical status data of the user is retrieved and the difference is calculated with the current health data to obtain the status change value. This method is similar to the feature extraction technology commonly used in medical data analysis, such as extracting health indicator features from historical data; judging the health status of the user according to the positive and negative zero of the status change value, which is similar to the classification method in health status assessment; real-time monitoring of the health status changes of the user, timely discovery of recovery, deterioration or stable state; this embodiment can provide early warning information to the user through the judgment of the status change value to help the user take measures in time. Step S2012 sums the status change value and the personal attribute value to obtain the patient's health condition. This method is similar to the comprehensive scoring method in health assessment; the patient's health status is analyzed to obtain the patient's risk value. This is similar to the risk score in the health risk assessment model; weight values ​​are assigned to relevant health data according to the patient's risk value, which is similar to feature weighting in data mining; this embodiment provides personalized health assessment through the combination of personal attribute values ​​and status change values, which is similar to health management based on individual differences; and through the calculation of risk values, personalized health advice and early warnings can be provided to patients to help users better manage their health. In step S2013, the data retrieval strategy and target retrieval strategy of the physiological knowledge model are retrieved based on the weight value; the data retrieval strategy is determined to be a paging retrieval strategy or a non-paging retrieval strategy; this embodiment uses a data retrieval strategy driven by a weight value to more efficiently manage a large amount of health data, reduce server burden, and improve data query efficiency; according to different data retrieval strategies, different target applications are supported to improve the flexibility and applicability of the system.

[0104] Embodiment 5:

[0105] like Figure 5 As shown, based on Example 3, the process of retrieving target data from the target retrieval database in step S202 provided in the embodiment of the present invention includes the following steps:

[0106] Step S2021: if it is determined that the data retrieval strategy is a paging retrieval strategy, then determine whether there is an address selection statement in the first call statement; if it is determined that there is a restriction condition statement in the first call statement, then determine whether the number of data return rows in the restriction condition statement is greater than a preset threshold;

[0107] Step S2022: If it is greater than the preset threshold, determine the target application according to the application information, and retrieve the target retrieval data of the target application in the target retrieval database according to the target application and the database policy; if the data retrieval policy is determined to be a non-paged retrieval policy, retrieve the retrieval data of the target application in the target retrieval database according to the data query volume of the target application;

[0108] Step S2023: if it is determined that the first calling statement does not contain any restrictive language, a preset restrictive statement is added to the first calling statement to obtain an updated second calling statement; the target database of the target application is called in the target calling database, and the called data is input into the user terminal.

[0109] In the above embodiment, in step S2021 of this embodiment, it is determined whether the data retrieval strategy is a paging retrieval strategy, and whether there is an addressing statement in the first call statement. If there is a restriction statement, it is further checked whether the number of data return rows exceeds the preset threshold; by checking the number of data return rows in the restriction statement, this embodiment can effectively control the scale of data query and avoid performance problems caused by returning too much data; the paging strategy can significantly improve performance when processing large amounts of data, especially when the amount of data is large, the performance of the paging method of the database service layer is more superior. Step S2022 determines the target application according to the application information. If it is a paging retrieval strategy and the number of returned rows exceeds the threshold, the data of the target application is retrieved in the target retrieval database. If it is a non-paging retrieval strategy, data is retrieved according to the data query volume of the target application; by dynamically adjusting the data retrieval strategy, this embodiment can optimize the data query process according to actual needs and improve the response speed and resource utilization of the system; the paging strategy can effectively reduce the amount of data in a single query, thereby reducing the burden on the server and improving query efficiency. In step S2023, if there is no restriction statement in the first call statement, a preset restriction statement is added thereto; the updated second call statement is used to retrieve data from the target call database and input the result to the user terminal; adding the restriction statement can better control the amount of data returned and avoid unnecessary data transmission, thereby improving system performance.

[0110] Embodiment 6:

[0111] like Figure 6 As shown, based on Example 3, the process of forming a personal data set in step S203 provided in this embodiment of the present invention includes the following steps:

[0112] Step S2031: Determine whether the preset application whitelist in the target database can match the target application; if it is determined that the target application can be matched in the preset application whitelist, determine the maximum number of return rows in the preset application whitelist; retrieve the target retrieval data of the target application in the target retrieval database according to the maximum number of return rows;

[0113] Step S2032: If the retrieval fails, it is determined as a calibration failure; the relevant health data is integrated with the historical relevant data and the comparative analysis results to form a personal data set;

[0114] Step S2033: performing preprocessing operations on the personal data set, such as removing outliers, balancing data distribution, converting categorical variables, and standardizing numerical variables; dividing the personal data set into a training set, a validation set, and a test set.

[0115] In the above embodiment, in step S2031 of this embodiment, by calling the preset application whitelist in the target database, check whether the target application can be matched; if the target application can be matched, the maximum return row number of the preset application whitelist is determined; according to the maximum return row number, the target retrieval data of the target application is retrieved in the target retrieval database; this embodiment ensures that only qualified application data is retrieved by matching the preset application whitelist, reducing the interference of irrelevant data; by determining the maximum return row number, the scope and efficiency of data retrieval can be effectively controlled to avoid unnecessary data processing. Step S2032 If the target retrieval data fails, it is determined as a calibration failure; the relevant health data is integrated with the historical relevant data and the comparative analysis results to form a personal data set; in this embodiment, through the determination of retrieval failure, problems in the data acquisition process are promptly identified and processed to avoid interruption of the process; multi-source data are integrated to build a more comprehensive personal data set to provide richer information for data analysis and model training. In step S2033, the personal data set is preprocessed by removing outliers, balancing data distribution, converting categorical variables, and standardizing numerical variables; the personal data set is divided into a training set, a validation set, and a test set; this embodiment improves the quality and consistency of data and reduces the impact of noise in model training by removing outliers and balancing data distribution. The model training optimization ensures the consistency of data format and improves the efficiency and accuracy of model training by converting categorical variables and standardizing numerical variables; the model performance is evaluated by dividing the training set, validation set, and test set, which can effectively evaluate the performance and generalization ability of the model and avoid overfitting or underfitting problems.

[0116] Embodiment 7:

[0117] like Figure 7As shown, based on Example 1, the process of inputting the simulation result into the diseased state simulation physiological knowledge data model in step S300 provided in the embodiment of the present invention includes the following steps:

[0118] Step S301: Retrieving historical disease data in the patient's physiological database based on the personal data set; using the training set in the personal data set to train the disease state model; during the training process, the personal data set is analyzed with the user's historical disease data to obtain the disease data trend;

[0119] Step S302: learning the disease data and using recursion to screen variables; based on each training result of the disease state model, gradually eliminating variables that are less than a preset value, and extracting the variables that contribute most to the disease state simulation result according to the final training disease state model simulation result;

[0120] Step S303: input the real-time physiological parameters into the disease state simulation model for simulation; input the simulation results into the physiological knowledge data model; the physiological knowledge data model outputs the simulation results to the user terminal; and find the corresponding doctor for answer based on the simulation results, and store the answer results.

[0121] In the above embodiment, in step S301 of this embodiment, historical disease data is retrieved from the patient's physiological database and preprocessed, including data cleaning, missing value processing and standardization; the disease state model is trained using the training set in the personal data set, which usually involves selecting appropriate features, feature extraction and feature engineering; trend analysis obtains the trend of disease data by analyzing the personal data set and the user's historical disease data; in this embodiment, the model can better capture the patient's health trend by using historical data for training; medical decision support can provide personalized health advice and early warning information to the patient by analyzing the patient's personal data; in step S302, a recursive method is used to screen variables and gradually eliminate unimportant variables; based on each training result, the model is gradually optimized to ensure the model's predictive performance; finally, the variables that contribute most to the disease state simulation results are extracted, which may involve feature importance ranking and influencing factor analysis; in this embodiment, recursive screening is used to reduce redundant variables in the model and improve the operating efficiency and interpretability of the model; and by extracting key variables, the predictive ability and accuracy of the model are enhanced. In step S303, the real-time physiological parameters are input into the disease state simulation model for simulation, which may involve real-time monitoring and data collection of multiple physiological parameters; the simulation results are input into the physiological knowledge data model, and the results are output through the user terminal; the corresponding physician is found for answers based on the simulation results, and the answers are stored; in this embodiment, through real-time data input and simulation, the system can timely monitor the patient's health status and provide early warning information; and through expert answers and storage mechanisms, the system can help medical institutions optimize resource allocation and improve the level of medical services.

[0122] Embodiment 8:

[0123] like Figure 8 As shown, based on Example 7, the process of extracting the variable that contributes most to the disease state simulation result in step S302 provided in the embodiment of the present invention includes the following steps:

[0124] Step S3021: Collect the node set of the training set and the user's historical disease data in the personal data set, analyze the output data and output data characteristics of each node in the node set, and analyze the dependency relationship between the nodes based on the characteristic analysis results of the node input data and output data;

[0125] Step S3022: constructing a dependency matrix between nodes according to the dependency relationship, calculating the similarity between the output data of the historical diseased node and the output data of the real-time health data node, and judging the dependency relationship between the output data of the historical diseased node and the data of the real-time health data node according to the threshold value between the similarities;

[0126] Among them, the formula for calculating the similarity between the output data of the historical diseased node and the output data of the real-time healthy data node; let S(h,r) be the similarity between the output data vector h of the historical diseased node and the output data vector r of the real-time healthy data node, the formula is as follows:

[0127]

[0128] Where h represents the output data vector of the historical diseased node, h = [h1,h2,…,h n ]; r represents the output data vector of the real-time health data node, r = [r1, r2, ..., r n ]; n represents the dimension of the data vector; w i represents the weight of the i-th data dimension, which is used to adjust the impact of different dimensions on similarity values; TF(h i ,r i ) represents the word frequency of the i-th dimension, indicating the importance of this dimension in the data;

[0129]

[0130] In the formula, IDF(h i ,r i ) represents the inverse document frequency of the i-th dimension, indicating the rarity of this dimension in the global data;

[0131]

[0132] Where N is the total number of data sets, II(h i =r j ) is the indicator function, when h i =r j 1 when it is, otherwise 0;

[0133] DTW(h i ,r i ) represents the dynamic time warping similarity of the time series in the i-th dimension, which is used to measure the similarity of time series data;

[0134]

[0135] Where τ is the time warping function and T is the length of the time series;

[0136] Cov(h i ,r i ) represents the covariance of the i-th dimension, which is used to measure the linear relationship between two data sequences;

[0137]

[0138] In the formula, μ h and μ r They are h i and r i The formula comprehensively considers the word frequency, inverse document frequency, dynamic time warping of the time series and covariance of the data, and calculates the final similarity value by weighted summation; the weight can be adjusted according to the specific application scenario to adapt to different data characteristics and needs; it can effectively calculate the similarity between historical disease nodes and real-time health data nodes;

[0139] Step S3023: Identify optimizable node configuration information and potential parallel execution sets and sequential execution sets between nodes based on the dependency relationships; and reorder and optimize the dependency configuration execution order based on the identified parallel execution sets and sequential execution sets.

[0140] In the above embodiment, step S3021 in this embodiment extracts the training set and the user's historical disease data from the personal data set to form a node set; the output data and its characteristics of each node are analyzed in detail to identify the dependencies between the nodes; this embodiment determines the dependencies between the nodes through feature analysis of the input and output data; the identification of dependencies helps to better understand the correlation between the data, thereby improving the depth and breadth of data analysis. Step S3022 constructs a dependency matrix based on the dependencies between the nodes, calculates the output data similarity between the historical disease nodes and the real-time health data nodes; calculates the output data similarity between the historical disease nodes and the real-time health data nodes; this embodiment can monitor the user's health status in real time and discover potential health problems in a timely manner by calculating the similarity; according to the similarity results, dynamically adjust the health management strategy to improve the pertinence and effectiveness of health management. Step S3033 identifies and optimizes node configuration information to improve the overall performance of the system, identifies the node sets that can be executed in parallel and the node sets that need to be executed sequentially, and reorders and optimizes the execution order of the dependency configuration based on the parallel and sequential execution sets; this embodiment can significantly improve the operating efficiency and response speed of the system by optimizing the node configuration and execution order; the identification of parallel execution sets enables the system to better handle the needs of multi-task parallel processing and enhance the flexibility and scalability of the system.

[0141] Embodiment 9:

[0142] like Fig. 9 As shown, based on Example 7, the process of outputting the simulation result to the user terminal in step S303 provided in the embodiment of the present invention includes the following steps:

[0143] Step S3031: based on each training result of the diseased state simulation model, gradually eliminate variables that are less than the threshold, and extract the variable value that contributes the most to the simulation result according to the final training diseased state simulation result; and obtain the final diseased state simulation model;

[0144] Step S3032: input the real-time physiological parameters into the disease state simulation model, simulate the real-time physiological parameters, and obtain the user's health data changes; compare the user's health data changes with the preset health data changes;

[0145] Step S3033: Evaluate the comparison result, input the evaluation result into the physiological knowledge data model; output the evaluation result to the user terminal, and find the corresponding physician for answer according to the evaluation result, and store the answer process.

[0146] In the above embodiment, step S3031 of this embodiment ensures that only variables that contribute significantly to the simulation results are retained in the model by gradually eliminating variables that are less than the threshold; extract key features from the training data, which are crucial to the predictive ability of the model; this embodiment retains the variables that contribute most to the simulation results, so that the model can more accurately reflect the actual disease state, thereby improving the accuracy of the prediction; eliminating irrelevant or less influential variables can reduce the risk of overfitting the model and make the model more generalizable on new data. Step S3032 uses the latest physiological parameter data for simulation to ensure that the health data changes output by the model are based on the latest user status; compare the simulation results with the preset health data changes to evaluate whether the user's health status is within the normal range; through real-time simulation and comparative analysis, this embodiment can timely detect the user's health abnormalities and provide early warnings and intervention suggestions for the user. According to the user's real-time health data changes, the system can provide personalized health management plans to help users better maintain their own health. Step S3033 performs a detailed evaluation of the comparison results, and feeds the evaluation results back to the physiological knowledge data model for model optimization; the evaluation results are output to the user terminal, and the corresponding physician is searched for answers based on the evaluation results, so that the user can obtain professional medical advice in a timely manner; in this embodiment, through timely feedback and professional answers, the user can obtain more comprehensive and accurate health management information, and improve the user's satisfaction and trust in the system; searching for corresponding physicians for answers based on the evaluation results helps to optimize the allocation of medical resources and ensure that users can obtain timely and effective medical services.

[0147] Embodiment 10:

[0148] like Fig.10 As shown, based on Embodiments 1 to 9, the smart medical interaction system based on big data provided by the embodiment of the present invention includes:

[0149] The health data collection module 1 is used to create a physiological knowledge data model based on the historical physiological knowledge database. The user enters personal attributes through the terminal to register, and different permissions are granted according to the personal attributes; the user enters relevant health data through the terminal, collects the patient's physiological data, and transmits the collected patient's physiological data to the physiological knowledge data model;

[0150] Among them, different permissions include patient permissions, doctor permissions, and management permissions;

[0151] Retrieving health data module 2, used to analyze relevant health data, and assign weight values ​​to real-time physiological parameters according to the analysis results; calibrating the physiological knowledge data model according to the weight values, and directly retrieving it if the calibration is successful; if the calibration fails, forming a personal data set; dividing the personal data set into a training set, a validation set, and a test set;

[0152] Among them, physiological data includes heart rate, blood pressure, blood sugar level, and medication intake;

[0153] The interactive health data module 3 is used to build a disease state simulation model based on the patient's physiological database, and use the training set to train the disease state simulation model; input the real-time physiological parameters into the disease state simulation model for simulation; input the simulation results into the disease state simulation physiological knowledge data model, find the corresponding doctor for answer according to the simulation results, and store the answer process.

[0154] In the above embodiment, the user of the health data collection module of this embodiment enters personal attributes through the terminal to register, and different permissions are granted according to personal attributes, including patient permissions, doctor permissions, and management permissions. This design ensures the security and privacy of the data, while meeting the needs of different roles; the user enters relevant health data such as heart rate, blood pressure, blood sugar level, and drug intake through the terminal, and the data is collected and transmitted to the physiological knowledge data model in real time; based on the user's personal attributes and health data, the system can provide personalized health management suggestions and early warning information, thereby improving the user's health management efficiency. The health data module is retrieved to analyze the relevant health data, and weight values ​​are assigned to real-time physiological parameters based on the analysis results. The physiological knowledge data model is calibrated according to the weight value, and if the calibration is successful, it is directly retrieved; if the calibration fails, a personal data set is formed; the personal data set is divided into a training set, a validation set and a test set for training and verification of the diseased state simulation model; through weight assignment and model calibration, the system in this embodiment can more accurately evaluate the user's health status and provide personalized health advice and early warning information; through the division and training of the data set, the diseased state simulation model is continuously optimized to improve the accuracy and reliability of the model; the interactive health data module constructs a diseased state simulation model based on the patient's physiological database and uses the training set for training; real-time physiological parameters are input into the diseased state simulation model for simulation; according to the simulation results, the corresponding doctor is found for answer, and the answer process is stored; through the diseased state simulation, the system in this embodiment can provide users with more intelligent health management services to help users better understand their own health status; the answer process is stored for users to check and trace at any time, which enhances the transparency and traceability of the system.

[0155] Fig.11 A block diagram of an exemplary electronic device suitable for implementing embodiments of the present invention is shown.

[0156] The electronic device may include a central processing unit / microprocessor / main control chip, etc. 4; a storage medium 5, coupled to the central processing unit / microprocessor / main control chip, etc. 4, and storing computer executable instructions therein, for performing the steps of each method of an embodiment of the present invention when executed by the processor.

[0157] The central processing unit / microprocessor / main control chip 4 may include but is not limited to, for example, one or more processors or microprocessors.

[0158] The storage medium 5 may include, but is not limited to, for example, random access memory (RAM), read-only memory (ROM), flash memory, EPROM memory, EEPROM memory, registers, computer storage media (such as hard disk, floppy disk, solid state drive, removable disk, CD-ROM, DVD-ROM, Blu-ray disc, etc.).

[0159] In addition, the electronic device may also include (but not limited to) a data bus 6, an input / output bus / external bus / device bus 7, a display 8, and input / output devices 9 (eg, keyboard, mouse, speaker, etc.).

[0160] The central processing unit / microprocessor / main control chip etc. 4 can communicate with external devices ( 8 , 9 etc.) through the I / O bus 7 via a wired or wireless network (not shown).

[0161] The storage medium 5 may also store at least one computer executable instruction for executing the various functions and / or method steps in the embodiments described in the present technology when executed by the central processing unit / microprocessor / main control chip 4.

[0162] In one embodiment, the at least one computer executable instruction may also be compiled into or constitute a software product, wherein one or more computer executable instructions are executed by a processor to perform the various functions and / or method steps in the embodiments described in the present technology.

[0163] Fig.12 A schematic diagram of a computer-readable storage medium according to an embodiment of the present invention is shown.

[0164] like Fig.12 As shown, instructions are stored on a non-transitory computer-readable storage medium 11, and the instructions are, for example, computer-readable instructions 10. When the computer-readable instructions 10 are executed by the processor, the various methods described above can be executed. Non-transitory computer-readable storage media include, but are not limited to, for example, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and / or cache memory (cache), etc. Non-transitory non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. For example, the non-transitory computer-readable storage medium 11 can be connected to a computing device such as a computer, and then, when the computing device runs the computer-readable instructions 10 stored on the computer-readable storage medium 11, the various methods described above can be performed.

[0165] In the several embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0166] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0167] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0168] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, 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, and the computer software product is stored in a storage medium, including a number of instructions for executing all or part of the steps of the various embodiments of the method of the present invention through a computer device (which can be a personal computer, a server, or a network device, etc.). The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (full name in English: Read-Only Memory, English abbreviation: ROM), random access memory (full name in English: Random Access Memory, English abbreviation: RAM), disk or optical disk and other media that can store program codes.

[0169] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A smart medical interaction method based on big data, characterized in that: The big data-based intelligent medical interaction method comprises the following steps: A physiological knowledge data model is created based on the historical physiological knowledge database. Users enter personal attributes through the terminal to register and are granted different permissions based on their personal attributes. Users enter relevant health data through the terminal to collect patient physiological data, and the collected patient physiological data is transferred to the physiological knowledge data model. Analyze relevant health data and assign weight values ​​to real-time physiological parameters based on the analysis results; The physiological knowledge data model is calibrated according to the weight value, and if the calibration is successful, it is directly retrieved; If the calibration fails, a personal data set is formed; Divide your personal data set into training, validation, and test sets; Building a disease state simulation model based on the patient's physiological database, and using the training set to train the disease state simulation model; The real-time physiological parameters are input into the disease state simulation model for simulation; the simulation results are input into the disease state simulation physiological knowledge data model, the corresponding physician is searched for answers according to the simulation results, and the answer process is stored.

2. The big data-based intelligent medical interaction method according to claim 1, characterized in that: The process of collecting patient physiological data includes the following steps: Create a physiological knowledge data model based on the historical physiological knowledge database, identify and connect various health data sources related to the research; obtain historical data from the selected data source; integrate different historical data into different physiological knowledge sub-datasets, and merge different physiological knowledge sub-datasets into a physiological knowledge dataset; The user registers by inputting personal attributes through the terminal, and is granted different permissions according to the personal attributes; the user inputs relevant health data through the terminal, collects the patient's physiological data, and transmits the collected patient's physiological data to the physiological knowledge data model; Clean, convert and format patient physiological data, and handle missing values ​​and outliers; The preprocessed patient physiological data is input into the physiological knowledge data model for screening.

3. The big data-based intelligent medical interaction method according to claim 1, characterized in that: The process of calibrating the physiological knowledge data model according to the weight value includes the following steps: Compare and analyze relevant health data with historical user data to obtain the patient's health status; assign weight values ​​based on the patient's health status, and retrieve the data retrieval strategy and target retrieval database of the physiological knowledge data model according to the weight values; Determine a database strategy of a target access database; retrieve target access data from the target access database based on the data access strategy and the database strategy; Output the target retrieved data to the user terminal; If the retrieval fails, the calibration is judged to have failed; the relevant health data is integrated with the historical relevant data and the comparative analysis results to form a personal data set.

4. The big data-based intelligent medical interaction method according to claim 3, characterized in that: The process of obtaining the patient's health status includes the following steps: Retrieve the user's historical status data, perform difference calculation with relevant health data, and obtain the status change value. If the status change value is positive, it is judged to be in a recovery state; if the status change value is negative, it is judged to be in a deterioration state; if the status change value is zero, it is judged to be in a stable state; The state change value and the personal attribute value are summed up to calculate the patient's health status; the patient's health status is analyzed to obtain the patient's risk value; Assign weight values ​​to relevant health data based on the patient's risk value; Data retrieval strategy and target retrieval strategy for retrieving physiological knowledge models based on weight values; Determine whether the data retrieval strategy is a paging retrieval strategy or a non-paging retrieval strategy, and determine the corresponding target application.

5. The big data-based intelligent medical interaction method according to claim 3, characterized in that: The process of retrieving target data in the target retrieval database includes the following steps: If it is determined that the data retrieval strategy is a paging retrieval strategy, then determine whether there is an address selection statement in the first call statement; if it is determined that there is a restriction condition statement in the first call statement, then determine whether the number of data return rows in the restriction condition statement is greater than a preset threshold; If it is greater than a preset threshold, the target application is determined according to the application information, and the target retrieval data of the target application is retrieved from the target retrieval database according to the target application and the database policy; If it is determined that the data retrieval strategy is a non-paged retrieval strategy, then according to the data query volume of the target application, the retrieval data of the target application is retrieved from the target retrieval database; Determining that the first call statement does not contain a restrictive condition language, adding a preset restrictive statement to the first call statement to obtain an updated second call statement; The target database of the target application is called in the target calling database, and the called data is input into the user terminal.

6. The big data-based intelligent medical interaction method according to claim 3, characterized in that: The process of forming a personal data set includes the following steps: Determine whether the preset application whitelist in the target database can match the target application; if it is determined that the target application can be matched in the preset application whitelist, then determine the maximum number of rows returned by the preset application whitelist; According to the maximum number of returned rows, the target call data of the target application is called in the target call database; If the retrieval fails, it is considered as calibration failure; Integrate relevant health data with historical data and comparative analysis results to form a personal data set; The personal data set is used to remove outliers, balance data distribution, transform categorical variables, and standardize numerical variables; the personal data set is divided into training set, validation set, and test set.

7. The big data-based intelligent medical interaction method according to claim 1, characterized in that: The process of inputting the simulation results into the disease state simulation physiological knowledge data model includes the following steps: Retrieving historical disease data from the patient's physiological database based on the personal data set; using the training set in the personal data set to train the disease state model; During the training process, the personal data set is analyzed with the user’s historical disease data to derive disease data trends; The disease data is actually learned and recursion is used to screen variables; Based on each training result of the diseased state model, variables with values ​​less than a preset value are gradually eliminated, and the variables with the greatest contribution to the diseased state simulation results are extracted according to the simulation results of the final trained diseased state model; Inputting real-time physiological parameters into the disease state simulation model for simulation; inputting simulation results into the physiological knowledge data model; The physiological knowledge data model outputs the simulation results to the user terminal; And according to the simulation results, the corresponding doctor is found to answer the question and the answer result is stored.

8. The big data-based intelligent medical interaction method according to claim 7, characterized in that: The process of extracting the variables that contribute most to the disease state simulation results includes the following steps: Collect the node set of the training set and the user's historical disease data in the personal data set, analyze the output data and output data characteristics of each node in the node set, and analyze the dependency relationship between nodes based on the characteristic analysis results of the node input data and output data; According to the dependency relationship, a dependency matrix between nodes is constructed, and the similarity value between the output data of the historical diseased node and the output data of the real-time healthy data node is calculated. According to the threshold between the similarities, the dependency relationship between the output data of the historical diseased node and the data of the real-time healthy data node is determined; According to the dependency relationship, the node configuration information that can be optimized, as well as the potential parallel execution sets and sequential execution sets between the nodes are identified; according to the identified parallel execution sets and sequential execution sets, the execution order of the dependency configuration is reordered and optimized.

9. The big data-based intelligent medical interaction method according to claim 7, characterized in that: The process of outputting simulation results to the user terminal includes the following steps: Based on each training result of the disease state simulation model, the variables with a value less than the threshold are gradually eliminated, and the variable value with the greatest contribution to the simulation result is extracted according to the final training disease state simulation result; The final disease state simulation model is obtained; Input the real-time physiological parameters into the disease state simulation model, simulate the real-time physiological parameters, and obtain the user's health data changes; Compare the user's health data changes with the preset health data changes; Evaluate the comparison results, and input the evaluation results into the physiological knowledge data model; The evaluation results are output to the user terminal, and the corresponding doctor is found for answering according to the evaluation results, and the answering process is stored.

10. A smart medical interaction system based on big data, the smart medical interaction method based on big data according to any one of claims 1 to 9, characterized in that: The intelligent medical interaction system based on big data includes: The health data collection module is used to create a physiological knowledge data model based on the historical physiological knowledge database. Users enter personal attributes through the terminal to register and are granted different permissions based on their personal attributes. Users enter relevant health data through the terminal to collect patient physiological data, and the collected patient physiological data is transferred to the physiological knowledge data model. Retrieve the health data module to analyze the relevant health data and assign weight values ​​to the real-time physiological parameters according to the analysis results; The physiological knowledge data model is calibrated according to the weight value, and if the calibration is successful, it is directly retrieved; If the calibration fails, a personal data set is formed; Divide your personal data set into training, validation, and test sets; An interactive health data module is used to build a disease state simulation model based on the patient's physiological database and train the disease state simulation model using a training set; The real-time physiological parameters are input into the disease state simulation model for simulation; the simulation results are input into the disease state simulation physiological knowledge data model, the corresponding physician is searched for answers according to the simulation results, and the answer process is stored.

Images