Smart medical remote service method and system based on big data and AI algorithm

Abstract

The invention relates to the technical field of remote monitoring, and discloses a smart medical remote service method and system based on big data and an AI algorithm, and the method comprises the steps: obtaining a medical demand of a target patient, constructing a heterogeneous medical terminal of the target patient, collecting the multi-dimensional physiological data of the target patient through the heterogeneous medical terminal, and transmitting the multi-dimensional physiological data to the server; the relevance between the multi-dimensional physiological data and the medical requirements is analyzed; mining associated data of the multi-dimensional physiological data, and establishing a patient sub-map of the target patient; related substructures of the patient subgraphs in the evolutionary knowledge graph are analyzed, graph parameters of the related substructures are output, and the graph parameters comprise entity association features, time sequence evolution features and grade labels; and analyzing the remote service parameters of the target patient under the medical remote service to execute the medical service of the target patient. According to the invention, the accuracy of remote analysis of the patient can be improved.

Description

Technical Field

[0001] The present invention relates to a smart medical remote service method and system based on big data and AI algorithms, belonging to the field of remote monitoring technology. Background Art

[0002] Smart medical telemedicine provides patients with remote, intelligent healthcare services. This service model breaks the geographical limitations of traditional medical services, allowing patients to enjoy timely, professional, and personalized healthcare services regardless of their location.

[0003] Traditional smart medical remote services mainly rely on basic communication methods such as video calls and emails for remote consultations and advice. This method of data collection and transmission is often not real-time and comprehensive enough, and cannot solve the data silo problem in traditional methods. It also lacks in-depth analysis and intelligent decision-making capabilities, resulting in inaccurate remote analysis of patients. Summary of the Invention

[0004] The present invention provides a smart medical remote service method and system based on big data and AI algorithms, the main purpose of which is to improve the accuracy of remote analysis of patients.

[0005] To achieve the above objectives, the present invention provides a smart medical remote service method based on big data and AI algorithms, comprising: Obtaining the medical needs of a target patient, building a heterogeneous medical terminal for the target patient, collecting multidimensional physiological data of the target patient through the heterogeneous medical terminal, and analyzing the correlation between the multidimensional physiological data and the medical needs; mining associated data of the multidimensional physiological data through the association, and establishing a patient sub-graph of the target patient based on the associated data and the multidimensional physiological data; Matching the patient subgraph with a relevant substructure of a pre-deployed evolvable knowledge graph, and outputting graph parameters of the relevant substructure, wherein the graph parameters include entity association features, temporal evolution features, and level labels; According to the atlas parameters, remote service parameters of the target patient under medical remote service are analyzed to perform medical service for the target patient.

[0006] Optionally, the step of obtaining the medical needs of the target patient includes: Collecting data sources under medical teleservices, wherein the data sources include structured data and semi-structured data, and wherein the data sources come from medical institutions and external public databases; Defining the core ontology and attribute expansion coefficients of the data source; According to the core ontology and attribute expansion coefficient, wherein the core ontology refers to the entity type in the knowledge graph, and the attribute expansion coefficient refers to the parameter used to dynamically adjust the weight or confidence of the entity attribute; Extracting entities and entity relationships in the data source based on the core ontology and attribute expansion coefficients; Storing the entities and entity relationships to obtain an initial knowledge graph for the medical teleservice; Determine the triggering conditions, update method and conflict handling process of the initial knowledge graph; The trigger conditions, update methods and conflict handling procedures are integrated into the initial knowledge graph to obtain an evolvable knowledge graph under the medical remote service.

[0007] Optionally, the step of constructing a heterogeneous medical terminal for a target patient includes: Analyzing the medical application scenario of the target patient based on the medical needs of the target patient and determining the monitoring equipment for the target patient; Determine the device interface standard and edge computing node of the monitoring device, and establish the adaptive networking topology and multimodal interactive interface of the monitoring device; Based on the adaptive networking topology, multimodal interactive interface and monitoring equipment, a heterogeneous medical terminal for the target patient is constructed.

[0008] Optionally, analyzing the correlation between the multidimensional physiological data and the medical needs includes: Analyzing the data dimensions of the medical needs; Identifying recommendation weights for the data dimensions; Analyzing the multidimensional physiological data to correspond to the physiological state of the target patient; Based on the physiological state, optimizing the recommendation weight to obtain a target weight; Based on the data dimensions and target weights, the correlation between the multidimensional physiological data and the medical needs is analyzed.

[0009] Optionally, analyzing the correlation between the multidimensional physiological data and the medical needs based on the data dimensions and target weights includes: determining data dimension measurement values ​​of the data dimensions based on the multidimensional physiological data; Obtaining standard reference values ​​for the data dimensions; Based on the data dimension measurements, the standard reference values, and the target weights, the following formula is used to analyze the correlation between the multidimensional physiological data and the medical needs: ; Wherein, R represents the correlation between multidimensional physiological data and the medical needs, represents the target weight of the cth data dimension of medical demand, represents the data dimension measurement value of the cth data dimension of medical demand, The standard reference value of the cth data dimension of medical needs, It represents the cosine similarity between the measured value of the data dimension and the standard reference value, sim represents the similarity calculation function, and m represents the number of data dimensions.

[0010] Optionally, mining associated data of the multidimensional physiological data through the association includes: analyzing the dimensional integrity of the multidimensional physiological data according to the correlation; constructing a static association network of the multidimensional physiological data through the dimensional integrity; identifying dynamic correlation patterns in the multidimensional physiological data; mining potential correlation features of the multidimensional physiological data according to the static correlation network and the dynamic correlation pattern; Based on the potential correlation features, correlation data of the multi-dimensional physiological data is mined.

[0011] Optionally, analyzing the dimensional integrity of the multidimensional physiological data according to the correlation includes: identifying missing dimensions of the multidimensional physiological data based on the correlation; Calculating the information entropy of the missing dimension; According to the information entropy of the missing dimension, the dimensional integrity of the multidimensional physiological data is calculated using the following formula: ; Among them, DCI represents the dimensional integrity of multidimensional physiological data. represents the target weight of the c-th data dimension, represents the data dimension measurement value of the cth data dimension of medical demand, m represents the number of data dimensions, represents the indicator function, represents the information entropy of the missing dimension, Iogm represents the maximum information entropy of the missing dimension, Indicates missing dimensions, Represents the information entropy calculation function.

[0012] Optionally, matching the patient subgraph with a relevant substructure of a pre-deployed evolvable knowledge graph includes: identifying static relationships, dynamic relationships, and temporal relationships of the patient subgraph; Respectively analyzing the static relationship, dynamic relationship, and temporal relationship and the static matching nodes, dynamic matching nodes, and temporal matching nodes of the evolvable knowledge graph; The related substructure of the patient subgraph in the evolvable knowledge graph is determined through the static matching nodes, the dynamic matching nodes and the temporal matching nodes.

[0013] Optionally, analyzing the remote service parameters of the target patient under the medical remote service according to the atlas parameters includes: Analyzing the feature association strength of the entity association features corresponding to the graph parameters; Analyzing the current status of the target patient according to the feature association strength; Analyzing the potential risks of the current state based on the temporal evolution characteristics and level labels corresponding to the graph parameters; Establishing a three-level risk matrix for the target patient; Determining the warning level of the target patient based on the potential risks and the three-level risk matrix; Based on the current state, the potential risk, and the warning level, remote service parameters for the target patient under medical remote service are determined.

[0014] In order to solve the above problems, the present invention also provides a smart medical remote service system based on big data and AI algorithms, the system comprising: a correlation analysis module, configured to obtain the medical needs of a target patient, construct a heterogeneous medical terminal for the target patient, collect multidimensional physiological data of the target patient through the heterogeneous medical terminal, and analyze the correlation between the multidimensional physiological data and the medical needs; a patient sub-graph construction module, configured to mine associated data of the multi-dimensional physiological data through the association, and establish a patient sub-graph of the target patient based on the associated data and the multi-dimensional physiological data; A related substructure determination module is used to match the patient subgraph with the related substructure of the pre-deployed evolvable knowledge graph and output graph parameters of the related substructure, wherein the graph parameters include entity association features, temporal evolution features, and level labels; The patient remote service module is used to analyze the remote service parameters of the target patient under the medical remote service according to the atlas parameters, so as to perform the medical service for the target patient.

[0015] Compared with the problems described in the background technology, first of all, by obtaining the medical needs of target patients and building heterogeneous medical terminals, the multi-dimensional physiological data of patients can be fully collected. These data not only include traditional vital signs monitoring, such as electrocardiogram, blood pressure, blood sugar, etc., but also cover exercise, sleep and other data collected by wearable devices. By analyzing the correlation between these multi-dimensional physiological data and the patient's medical needs, the patient's health status and potential risks can be more accurately grasped. The correlation analysis not only reveals the intrinsic connection between the data, but also provides strong support for subsequent medical services. After mining the correlation data of multi-dimensional physiological data, a patient sub-graph of the target patient is established. The sub-graph not only contains the patient's static medical information, such as diagnosis results, Medication records, etc., also reflect the patient's dynamic physiological state in real time. By analyzing the relevant substructures of the patient's subgraph in the evolvable knowledge graph and outputting graph parameters such as entity association features, temporal evolution features, and level labels, it is possible to have a deeper understanding of the patient's disease evolution and risk trends. The graph parameters provide a scientific basis for medical decision-making, making medical services more accurate and personalized. The remote service parameters obtained based on the analysis of the graph parameters can provide customized medical services for patients. These service parameters include but are not limited to monitoring frequency, intervention measures, communication protocols, etc. By monitoring the patient's physiological data in real time and pushing emergency intervention suggestions when necessary, it can effectively prevent the disease from worsening and reduce unnecessary waste of medical resources. Therefore, the present invention can improve the accuracy of remote analysis of patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A flowchart of a smart medical remote service method based on big data and AI algorithms provided by one embodiment of the present invention; Figure 2 A schematic diagram of modules for implementing the smart medical remote service method based on big data and AI algorithms according to an embodiment of the present invention.

[0017] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0018] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0019] The present embodiment provides a method for intelligent medical remote service based on big data and AI algorithms. The execution entity of the method includes, but is not limited to, at least one of electronic devices such as a server or a terminal that can be configured to execute the method provided in the present embodiment. In other words, the method can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster.

[0020] Example 1

[0021] Reference Figure 1 FIG2 is a flow chart of a smart medical remote service method based on big data and AI algorithms according to an embodiment of the present invention. In this embodiment, the smart medical remote service method based on big data and AI algorithms includes: S1. Obtain the medical needs of a target patient, and construct a heterogeneous medical terminal for the target patient, collect multidimensional physiological data of the target patient through the heterogeneous medical terminal, and analyze the correlation between the multidimensional physiological data and the medical needs.

[0022] In detail, the steps before obtaining the target patient's medical needs include: Collecting data sources under medical teleservices, wherein the data sources include structured data and semi-structured data, and wherein the data sources come from medical institutions and external public databases; Defining the core ontology and attribute expansion coefficients of the data source; According to the core ontology and attribute expansion coefficient, wherein the core ontology refers to the entity type in the knowledge graph, and the attribute expansion coefficient refers to the parameter used to dynamically adjust the entity attribute weight; Extracting entities and entity relationships in the data source based on the core ontology and attribute expansion coefficients; Storing the entities and entity relationships to obtain an initial knowledge graph for the medical teleservice; Determine the triggering conditions, update method and conflict handling process of the initial knowledge graph; The trigger conditions, update methods and conflict handling procedures are integrated into the initial knowledge graph to obtain an evolvable knowledge graph under the medical remote service.

[0023] Among them, the structured data refers to standardized medical data with a fixed format that can be directly used for analysis, such as electronic medical records (diagnostic codes (ICD-10), medication records (RxNorm)), laboratory data (test results (LOINC codes), genetic testing reports (HL7 format)); the semi-structured data refers to medical data with partial organizational structure but requires further analysis, such as clinical guidelines (the hierarchical treatment plan in the NCCN guidelines (PDF / HTML format)); the core ontology refers to the entity type in the knowledge graph; the attribute expansion coefficient refers to the parameter used to dynamically adjust the entity attribute weight; the entity refers to the basic unit instance in the knowledge graph, such as disease entities and drug entities; the entity relationship refers to the directed edge connecting entities; the initial knowledge graph refers to the static knowledge network obtained after storing the entities and entity relationships; the evolvable knowledge graph refers to a dynamic knowledge network that can evolve under medical teleservices.

[0024] Optionally, the extracting of entities and entity relationships from the data source based on the core ontology and attribute expansion coefficients may be performed by performing entity recognition and relationship extraction through natural language processing (NLP) technology.

[0025] Optionally, the entities and entity relationships are stored to obtain an initial knowledge graph under the medical remote service, which can be stored in a graph database Neo4j.

[0026] Optionally, the trigger conditions include: new data input (e.g., a patient's latest test results) or external knowledge updates (e.g., a new drug indication added by the FDA); update methods include: adding or deleting corresponding nodes / relationships in the initial knowledge graph (e.g., adding "cardiovascular protective effects of SGLT2 inhibitors") or adjusting attribute weights (e.g., lowering the recommendation level for a particular drug based on recent research); and conflict resolution: when new and old evidence conflict, the more authoritative and updated source is prioritized. For example, the COVID-19 treatment graph automatically adds a "Paxil Indicators" node after the WHO updates its guidelines.

[0027] It should be explained that the target patients refer to specific individuals or groups who need to provide medical services, such as patients with specific diseases, people in specific age groups, people in specific regions, people with specific lifestyles, people with specific genotypes, etc. The medical needs refer to the problems, expectations and needs faced by the target patients in medical and health care, such as disease analysis, rehabilitation plans, health consultation and other needs.

[0028] The heterogeneous medical terminal constructed by the present invention for target patients is different from a single terminal solution and improves the integrity of data collection.

[0029] In detail, the construction of a heterogeneous medical terminal for a target patient includes: Analyzing the medical application scenario of the target patient based on the medical needs of the target patient and determining the monitoring equipment for the target patient; Determine the device interface standard and edge computing node of the monitoring device, and establish the adaptive networking topology and multimodal interactive interface of the monitoring device; Based on the adaptive networking topology, multimodal interactive interface and monitoring equipment, a heterogeneous medical terminal for the target patient is constructed.

[0030] Among them, the medical application scenario refers to the specific medical environment of the target patient when receiving telemedicine services, such as the physical scenario (home, community clinic, ambulance transfer, inpatient ward), disease stage (acute attack period, chronic stable period, rehabilitation training), the monitoring equipment refers to the hardware device used to collect the patient's physiological, biochemical or environmental data, such as vital signs (smart bracelet monitoring heart rate / blood oxygen), pathological indicators (blood glucose meter detecting fingertip blood glucose), the device interface standard refers to the technical protocol for data exchange between medical equipment and systems, including physical interface (USB, Bluetooth, Zigbee), data protocol and semantic standard, the edge computing node refers to the computing unit deployed near the monitoring equipment, the adaptive networking topology refers to the device connection architecture that is dynamically adjusted according to environmental changes, the multimodal interaction interface refers to the human-computer interaction interface that supports multiple natural methods, and the heterogeneous medical terminal refers to the complete patient-end system composed of the aforementioned elements.

[0031] Optionally, the adaptive networking topology and multimodal interaction interface of the monitoring device are established based on the device interface standard and edge computing nodes, wherein the adaptive networking topology can be achieved through wireless sensor network (WSN) technology: utilizing the self-organization and self-healing characteristics of wireless sensor networks to achieve adaptive networking of monitoring devices, and nodes can automatically adjust the network topology based on information such as signal strength and neighbor node status to adapt to environmental changes and device movement. The multimodal interaction interface can be constructed through natural user interface (NUI) technology: utilizing multiple interaction methods such as voice recognition, gesture recognition, and touch input.

[0032] It should be explained that the multidimensional physiological data refers to a collection of various types of biomedical data collected through heterogeneous medical terminals, covering the patient's full range of health status, including dynamic vital signs, imaging data, genomic data and environmental parameters.

[0033] The present invention analyzes the correlation between the multidimensional physiological data and the medical needs to determine the integrity of the target patient data, thereby improving the reliability of subsequent data analysis.

[0034] In detail, the analyzing the correlation between the multi-dimensional physiological data and the medical needs includes: Analyzing the data dimensions of the medical needs; Identifying recommendation weights for the data dimensions; Analyzing the multidimensional physiological data to correspond to the physiological state of the target patient; Based on the physiological state, optimizing the recommendation weight to obtain a target weight; Based on the data dimensions and target weights, the correlation between the multidimensional physiological data and the medical needs is analyzed.

[0035] Among them, the data dimensions refer to various physiological parameters and indicators involved in analyzing medical needs, such as vital signs (heart rate, blood pressure, body temperature, respiratory rate), biochemical indicators (blood sugar, blood lipids, liver function, kidney function); the recommended weight refers to the initial importance weight assigned to each data dimension based on the medical guideline analysis; the physiological state refers to the current physical health status and physiological function level of the target patient; the target weight refers to the final importance weight determined for each data dimension after comprehensively considering the recommended weight and optimized weight; the correlation refers to the degree of correlation between multidimensional physiological data and medical needs.

[0036] Optionally, the analyzing the multi-dimensional physiological data to determine the physiological state of the target patient can be performed by performing a correlation analysis between the multi-dimensional physiological data and the potential state of the target patient.

[0037] Furthermore, analyzing the correlation between the multidimensional physiological data and the medical needs based on the data dimensions and target weights includes: determining data dimension measurement values ​​of the data dimensions based on the multidimensional physiological data; Obtaining standard reference values ​​for the data dimensions; Based on the data dimension measurements, the standard reference values, and the target weights, the following formula is used to analyze the correlation between the multidimensional physiological data and the medical needs: ; Wherein, R represents the correlation between multidimensional physiological data and the medical needs, represents the target weight of the cth data dimension of medical demand, represents the data dimension measurement value of the cth data dimension of medical demand, The standard reference value of the cth data dimension of medical needs, It represents the cosine similarity between the measured value of the data dimension and the standard reference value, Sim represents the similarity calculation function, and m represents the number of data dimensions.

[0038] The data dimension measurement value refers to the physiological data value actually collected from the target patient for each data dimension. For example, if the data dimension is "blood pressure", the data dimension measurement value may be "120 / 80 mmHg", and the standard reference value refers to the normal or expected value range recognized by medicine or determined based on a large amount of statistical data for each data dimension. Taking "blood pressure" as an example, the standard reference value may be "90-120 / 60-80 mmHg", and the cosine similarity refers to the similarity between the data dimension measurement value and the standard reference value, and the similarity calculation function refers to the function used to calculate the similarity between the data dimension measurement value and the standard reference value.

[0039] S2. Mining the associated data of the multidimensional physiological data through the association, and establishing a patient sub-graph of the target patient based on the associated data and the multidimensional physiological data.

[0040] The present invention uses the correlation to mine the associated data of the multidimensional physiological data to supplement the relevant data and improve the accuracy of the subsequent patient map construction.

[0041] In detail, mining the associated data of the multi-dimensional physiological data through the association includes: analyzing the dimensional integrity of the multidimensional physiological data according to the correlation; constructing a static association network of the multidimensional physiological data through the dimensional integrity; identifying dynamic correlation patterns in the multidimensional physiological data; mining potential correlation features of the multidimensional physiological data according to the static correlation network and the dynamic correlation pattern; Based on the potential correlation features, correlation data of the multidimensional physiological data is mined.

[0042] Dimensional integrity refers to the degree to which each data dimension in a given multidimensional physiological data set fully and accurately reflects the information it should contain. The static association network refers to a network model that reflects the relationships between data dimensions, constructed based on the correlations between data dimensions. The dynamic association pattern refers to how the interactions between certain physiological indicators change over time, environment, or therapeutic intervention in multidimensional physiological data analysis. For example, the relationship between heart rate and blood pressure may be different after exercise than at rest. The potential association feature refers to a potential association feature in multidimensional physiological data that may be manifested as a certain physiological state or disease risk implied by the joint changes of multiple physiological indicators. For example, simultaneous abnormalities in multiple indicators may indicate the early stages of a disease. The associated data refers to a set of data directly related to a specific goal or problem identified based on the potential association feature. For example, in heart disease analysis, the associated data may include indicators such as electrocardiogram, blood pressure, and blood lipids.

[0043] Optionally, the static association network of the multidimensional physiological data is constructed through the dimensional integrity. An undirected or directed graph structure is constructed through the dimensional integrity, and the graph structure is optimized using a graph theory algorithm (minimum spanning tree (MST)) to generate the static association network of the multidimensional physiological data.

[0044] Furthermore, analyzing the dimensional integrity of the multidimensional physiological data according to the correlation includes: identifying missing dimensions of the multidimensional physiological data based on the correlation; Calculating the information entropy of the missing dimension; According to the information entropy of the missing dimension, the dimensional integrity of the multidimensional physiological data is calculated using the following formula: ; Among them, DCI represents the dimensional integrity of multidimensional physiological data. represents the target weight of the c-th data dimension, represents the data dimension measurement value of the cth data dimension of medical demand, m represents the number of data dimensions, represents the indicator function, represents the information entropy of the missing dimension, Iogm represents the maximum information entropy of the missing dimension, Indicates missing dimensions, Represents the information entropy calculation function.

[0045] Among them, the missing dimension refers to the data dimension that cannot be collected in the multidimensional physiological data set due to various reasons (such as equipment failure, data collection omission, etc.), the information entropy refers to the function used to quantify the degree of confusion of a certain data dimension, the indicator function refers to the function that is 1 when the data dimension is successfully collected and 0 otherwise, the maximum information entropy refers to the maximum uncertainty that a random variable may reach under given conditions, and the information entropy calculation function refers to the mathematical function used to calculate the information entropy of random variables.

[0046] The present invention establishes a patient subgraph for the target patient based on the associated data and the multidimensional physiological data, which can be connected to the evolvable knowledge graph, providing a basis for analyzing the remote service parameters of the target patient. The patient subgraph refers to a graph related to the target patient extracted from the entire multidimensional physiological data network.

[0047] S3. Match the patient subgraph with the relevant substructure of the pre-deployed evolvable knowledge graph, and output the graph parameters of the relevant substructure, wherein the graph parameters include entity association features, temporal evolution features, and level labels.

[0048] Analyzing the relevant substructure of the patient subgraph in the evolvable knowledge graph can effectively analyze the patient subgraph in the relevant substructure of the evolvable knowledge graph, which can improve the accuracy of the analysis of the target patient's status.

[0049] In detail, matching the patient subgraph with the relevant substructure of the pre-deployed evolvable knowledge graph includes: Identifying static relationships, dynamic relationships, and temporal relationships of the patient subgraphs; Respectively analyzing the static relationship, dynamic relationship, and temporal relationship and the static matching nodes, dynamic matching nodes, and temporal matching nodes of the evolvable knowledge graph; The related substructure of the patient subgraph in the evolvable knowledge graph is determined through the static matching nodes, the dynamic matching nodes and the temporal matching nodes.

[0050] Among them, the static relationship refers to a relationship that does not change with time, such as the diagnostic relationship between the target patient and the disease (such as "target patient A suffers from hypertension"); the dynamic relationship refers to a relationship that changes with time, such as the adjustment of the target patient's medication dosage or the change of the condition; the temporal relationship refers to a relationship constructed according to the time attribute, for example, "On January 1, 2024, the blood glucose value of patient A was 7.0 mmol / L", the static matching node refers to the node in the evolvable knowledge graph corresponding to the static relationship with the patient subgraph, the dynamic matching node refers to the node in the evolvable knowledge graph corresponding to the dynamic relationship with the patient subgraph, the temporal matching node refers to the node in the evolvable knowledge graph corresponding to the temporal relationship with the patient subgraph, and the related substructure refers to the part of the structure in the evolvable knowledge graph that matches the static relationship, dynamic relationship and temporal relationship in the patient subgraph.

[0051] Optionally, the respectively analyzing the static relationship, dynamic relationship and temporal relationship and the static matching nodes, dynamic matching nodes and temporal matching nodes of the evolvable knowledge graph may be achieved by a matching algorithm.

[0052] It should be explained that the entity association features refer to the description of the association patterns between entities in related substructures (such as diseases, symptoms, physiological indicators), including direct connections, indirect paths, co-occurrence frequencies, etc. The temporal evolution features refer to the description of the changing patterns of entity associations in related substructures over time, including dynamic relationship strength, pattern stability within the time window, etc. The hierarchical labels refer to the hierarchical classification of entities and relationships in related substructures to reflect their importance in the knowledge graph.

[0053] S4. Analyze the remote service parameters of the target patient under the medical remote service according to the atlas parameters to perform the medical service for the target patient.

[0054] The present invention analyzes the remote service parameters of the target patient under medical remote service according to the atlas parameters to perform medical service for the target patient, thereby realizing efficient medical remote service for the target patient.

[0055] In detail, analyzing the remote service parameters of the target patient under the medical remote service according to the atlas parameters includes: Analyzing the feature association strength of the entity association features corresponding to the graph parameters; Analyzing the current status of the target patient according to the feature association strength; Analyzing the potential risks of the current state based on the temporal evolution characteristics and level labels corresponding to the graph parameters; Establishing a three-level risk matrix for the target patient; Determining the warning level of the target patient based on the potential risks and the three-level risk matrix; Based on the current state, the potential risk, and the warning level, remote service parameters for the target patient under medical remote service are determined.

[0056] Among them, the feature association strength refers to the degree of association between different entities in the graph parameters (such as symptoms, physiological indicators, diagnostic results, etc.); the current state refers to the current health status of the target patient obtained based on the feature association strength analysis; the potential risk refers to the prediction of the risk that the target patient may face in the future based on the time-series evolution characteristics and level labels in the graph parameters; the three-level risk matrix refers to a tool for assessing and classifying patient risks, usually including three levels (such as high, medium, and low); the warning level refers to the level determined by the potential risk and the three-level risk matrix, used to indicate the patient's current risk status; the remote service parameters refer to specific service parameters formulated for medical remote services based on the current state, potential risk and warning level. For example, for high-risk patients, remote service parameters may include monitoring physiological indicators once an hour and automatically pushing emergency intervention recommendations; for low-risk patients, remote service parameters may include monitoring physiological indicators once a day and sending health reminders regularly.

[0057] Optionally, the three-level risk matrix for the target patient can be constructed by analyzing the potential risk, the temporal trend of the temporal evolution characteristics, and the grade label. Exemplarily, the three-level risk matrix:

[0058] Compared with the problems described in the background technology, first of all, by obtaining the medical needs of target patients and building heterogeneous medical terminals, the multi-dimensional physiological data of patients can be fully collected. These data not only include traditional vital signs monitoring, such as electrocardiogram, blood pressure, blood sugar, etc., but also cover exercise, sleep and other data collected by wearable devices. By analyzing the correlation between these multi-dimensional physiological data and the patient's medical needs, the patient's health status and potential risks can be more accurately grasped. The correlation analysis not only reveals the intrinsic connection between the data, but also provides strong support for subsequent medical services. After mining the correlation data of multi-dimensional physiological data, a patient sub-graph of the target patient is established. The sub-graph not only contains the patient's static medical information, such as diagnosis results, Medication records, etc., also reflect the patient's dynamic physiological state in real time. By analyzing the relevant substructures of the patient's subgraph in the evolvable knowledge graph and outputting graph parameters such as entity association features, temporal evolution features, and level labels, it is possible to have a deeper understanding of the patient's disease evolution and risk trends. The graph parameters provide a scientific basis for medical decision-making, making medical services more accurate and personalized. The remote service parameters obtained based on the analysis of the graph parameters can provide customized medical services for patients. These service parameters include but are not limited to monitoring frequency, intervention measures, communication protocols, etc. By monitoring the patient's physiological data in real time and pushing emergency intervention suggestions when necessary, it can effectively prevent the disease from worsening and reduce unnecessary waste of medical resources. Therefore, the present invention can improve the accuracy of remote analysis of patients.

[0059] Example 2 like Figure 2 The figure shows a functional module diagram of a smart medical remote service system based on big data and AI algorithm according to the present invention.

[0060] The intelligent medical remote service system 200 based on big data and AI algorithms described in the present invention can be installed in an electronic device. Depending on the functionality implemented, the intelligent medical remote service system based on big data and AI algorithms can include a correlation analysis module 201, a patient sub-graph construction module 202, a related substructure determination module 203, and a patient remote service module 204. A module, also referred to as a unit, is a series of computer program segments that can be executed by an electronic device processor and perform a fixed function. These are stored in the electronic device's memory.

[0061] In the embodiment of the present invention, the functions of each module / unit are as follows: The correlation analysis module 201 is used to obtain the medical needs of the target patient, build a heterogeneous medical terminal for the target patient, collect multidimensional physiological data of the target patient through the heterogeneous medical terminal, and analyze the correlation between the multidimensional physiological data and the medical needs; The patient sub-graph construction module 202 is configured to mine the associated data of the multi-dimensional physiological data through the association, and to establish a patient sub-graph of the target patient based on the associated data and the multi-dimensional physiological data; The relevant substructure determination module 203 is used to match the relevant substructure of the patient subgraph with the pre-deployed evolvable knowledge graph and output graph parameters of the relevant substructure, wherein the graph parameters include entity association features, temporal evolution features, and level labels; The patient remote service module 204 is configured to analyze the remote service parameters of the target patient under the medical remote service according to the atlas parameters, so as to perform the medical service for the target patient.

[0062] In detail, each module in the smart medical remote service system 200 based on big data and AI algorithm in the embodiment of the present invention adopts the same Figure 1 The same technical means are used in the smart medical remote service method based on big data and AI algorithms described in the previous section, and can produce the same technical effects, so I will not go into details here.

[0063] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0064] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not limiting. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A smart medical remote service method based on big data and AI algorithm, characterized by: The method comprises: Obtaining the medical needs of a target patient, building a heterogeneous medical terminal for the target patient, collecting multidimensional physiological data of the target patient through the heterogeneous medical terminal, and analyzing the correlation between the multidimensional physiological data and the medical needs; mining associated data of the multidimensional physiological data through the association, and establishing a patient sub-graph of the target patient based on the associated data and the multidimensional physiological data; Matching the patient subgraph with a relevant substructure of a pre-deployed evolvable knowledge graph, and outputting graph parameters of the relevant substructure, wherein the graph parameters include entity association features, temporal evolution features, and level labels; According to the atlas parameters, remote service parameters of the target patient under medical remote service are analyzed to perform medical service for the target patient.

2. The smart medical remote service method based on big data and AI algorithm as claimed in claim 1, characterized in that: The steps before obtaining the target patient's medical needs include: Collecting data sources under medical teleservices, wherein the data sources include structured data and semi-structured data, and wherein the data sources come from medical institutions and external public databases; Defining the core ontology and attribute expansion coefficients of the data source; According to the core ontology and attribute expansion coefficient, wherein the core ontology refers to the entity type in the knowledge graph, and the attribute expansion coefficient refers to the parameter used to dynamically adjust the weight or confidence of the entity attribute; Extracting entities and entity relationships in the data source based on the core ontology and attribute expansion coefficients; Storing the entities and entity relationships to obtain an initial knowledge graph for the medical teleservice; Determine the triggering conditions, update method and conflict handling process of the initial knowledge graph; The trigger conditions, update methods and conflict handling procedures are integrated into the initial knowledge graph to obtain an evolvable knowledge graph under the medical remote service.

3. The smart medical remote service method based on big data and AI algorithm as claimed in claim 2, characterized in that: The construction of a heterogeneous medical terminal for a target patient includes: Analyzing the medical application scenario of the target patient based on the medical needs of the target patient and determining the monitoring equipment for the target patient; Determine the device interface standard and edge computing node of the monitoring device, and establish the adaptive networking topology and multimodal interactive interface of the monitoring device; Based on the adaptive networking topology, multimodal interactive interface and monitoring equipment, a heterogeneous medical terminal for the target patient is constructed.

4. The smart medical remote service method based on big data and AI algorithm as claimed in claim 3, characterized in that: The analyzing the correlation between the multi-dimensional physiological data and the medical needs includes: Analyzing the data dimensions of the medical needs; Identifying recommendation weights for the data dimensions; Analyzing the multidimensional physiological data to correspond to the physiological state of the target patient; Based on the physiological state, optimizing the recommendation weight to obtain a target weight; Based on the data dimensions and target weights, the correlation between the multidimensional physiological data and the medical needs is analyzed.

5. The smart medical remote service method based on big data and AI algorithm as claimed in claim 4, characterized in that: The analyzing the correlation between the multidimensional physiological data and the medical needs based on the data dimensions and the target weights includes: determining data dimension measurement values ​​of the data dimensions based on the multidimensional physiological data; Obtaining standard reference values ​​for the data dimensions; Based on the data dimension measurements, the standard reference values, and the target weights, the following formula is used to analyze the correlation between the multidimensional physiological data and the medical needs: ; Wherein, R represents the correlation between the multidimensional physiological data and the medical needs, represents the target weight of the cth data dimension of medical demand, represents the data dimension measurement value of the cth data dimension of medical demand, The standard reference value of the cth data dimension of medical needs, It represents the cosine similarity between the measured value of the data dimension and the standard reference value, Sim represents the similarity calculation function, and m represents the number of data dimensions.

6. The smart medical remote service method based on big data and AI algorithm as claimed in claim 5, characterized in that: The mining of associated data of the multi-dimensional physiological data through the association includes: analyzing the dimensional integrity of the multidimensional physiological data according to the correlation; constructing a static association network of the multidimensional physiological data through the dimensional integrity; identifying dynamic correlation patterns in the multidimensional physiological data; mining potential correlation features of the multidimensional physiological data according to the static correlation network and the dynamic correlation pattern; Based on the potential correlation features, correlation data of the multi-dimensional physiological data is mined.

7. The smart medical remote service method based on big data and AI algorithm according to claim 6, characterized in that: Analyzing the dimensional integrity of the multidimensional physiological data according to the correlation includes: identifying missing dimensions of the multidimensional physiological data based on the correlation; Calculating the information entropy of the missing dimension; According to the information entropy of the missing dimension, the dimensional integrity of the multidimensional physiological data is calculated using the following formula: ; Among them, DCI represents the dimensional integrity of multidimensional physiological data. represents the target weight of the c-th data dimension, represents the data dimension measurement value of the cth data dimension of medical demand, m represents the number of data dimensions, represents the indicator function, represents the information entropy of the missing dimension, Iogm represents the maximum information entropy of the missing dimension, Indicates missing dimensions, Represents the information entropy calculation function.

8. The smart medical remote service method based on big data and AI algorithm as claimed in claim 7, characterized in that: Matching the patient subgraph with a relevant substructure of a pre-deployed evolvable knowledge graph includes: Identifying static, dynamic, and temporal relationships of the patient subgraphs; Respectively analyzing the static relationship, dynamic relationship, and temporal relationship and the static matching nodes, dynamic matching nodes, and temporal matching nodes of the evolvable knowledge graph; The related substructure of the patient subgraph in the evolvable knowledge graph is determined through the static matching nodes, the dynamic matching nodes and the temporal matching nodes.

9. The smart medical remote service method based on big data and AI algorithm as claimed in claim 8, characterized in that: Analyzing the remote service parameters of the target patient under the medical remote service according to the atlas parameters includes: Analyzing the feature association strength of the entity association features corresponding to the graph parameters; Analyzing the current status of the target patient according to the feature association strength; Analyzing the potential risks of the current state based on the temporal evolution characteristics and level labels corresponding to the graph parameters; Establishing a three-level risk matrix for the target patient; Determining the warning level of the target patient based on the potential risks and the three-level risk matrix; Based on the current state, the potential risk, and the warning level, remote service parameters for the target patient under medical remote service are determined.

10. A smart medical remote service system based on big data and AI algorithm, characterized by: The system comprises: a correlation analysis module, configured to obtain the medical needs of a target patient, construct a heterogeneous medical terminal for the target patient, collect multidimensional physiological data of the target patient through the heterogeneous medical terminal, and analyze the correlation between the multidimensional physiological data and the medical needs; a patient sub-graph construction module, configured to mine associated data of the multi-dimensional physiological data through the association, and establish a patient sub-graph of the target patient based on the associated data and the multi-dimensional physiological data; A related substructure determination module is used to match the patient subgraph with the related substructure of the pre-deployed evolvable knowledge graph and output graph parameters of the related substructure, wherein the graph parameters include entity association features, temporal evolution features, and level labels; The patient remote service module is used to analyze the remote service parameters of the target patient under the medical remote service according to the atlas parameters, so as to perform the medical service for the target patient.

Images