Risk assessment system for chronic diseases

Abstract

The invention relates to the technical field of medical data processing, in particular to a risk assessment system for chronic diseases. The system is used for performing physiological index analysis according to a medical physiological monitoring signal to generate physiological index data; performing individual physiological index weight empowerment on the physiological index data based on genomics data to obtain individual physiological index data; performing quantitative analysis on pathological influence factors of chronic disease types according to the electronic health record data and the individual physiological index data, and generating individual chronic disease influence factor quantitative data; performing composite interaction risk factor analysis on the individual chronic disease influence factors based on the quantitative data of the individual chronic disease influence factors to generate individual chronic disease composite interaction risk factor data; and performing individual chronic disease risk assessment based on the individual chronic disease composite interaction risk factor data to generate individual chronic disease risk assessment data. According to the invention, chronic disease health management and accurate evaluation are realized through collaborative evaluation work of chronic disease risks.

Description

Technical Field

[0001] The present invention relates to the technical field of medical data processing, and in particular to a risk assessment system for chronic diseases. Background Art

[0002] Chronic diseases (such as cardiovascular disease, diabetes, chronic respiratory diseases, etc.) are one of the leading causes of death and disability worldwide. Their onset is characterized by long-term and hidden nature, and is usually caused by the combined effects of multiple biological, environmental and behavioral factors. With the development of modern medicine, the prevention, early diagnosis and individualized management of chronic diseases have gradually become research hotspots. Although medical technology has made significant progress in the diagnosis and treatment of chronic diseases, the means of early prevention and accurate risk assessment are still insufficient. Medical auxiliary tools based on artificial intelligence technology and big data analysis conduct a comprehensive and accurate assessment of individual chronic disease risks, assist medical professionals in discovering abnormal information about patients' chronic diseases earlier, and reduce the risk of chronic diseases, thereby promoting the development of health management in the direction of intelligence. However, the existing chronic disease risk assessment system performs chronic disease risk assessment based on electronic health records, which is difficult to fully reflect the health status of individuals. At the same time, the occurrence of chronic diseases is often caused by the interaction of multiple factors such as physiology and genetics. These complex relationships require more advanced analysis methods to explore potential pathological mechanisms and risk characteristics. Summary of the invention

[0003] Based on this, the present invention provides a risk assessment system for chronic diseases to solve at least one of the above technical problems.

[0004] To achieve the above purpose, a chronic disease risk assessment system includes the following modules: A multi-source physiological sign data collection module is used to collect multi-source physiological sign data of patients through a medical cloud service platform to generate multi-source physiological sign data, wherein the multi-source physiological sign data includes electronic health record data, medical physiological monitoring signals and genomics data; A physiological index analysis module is used to analyze physiological indexes according to medical physiological monitoring signals and generate physiological index data; The physiological index potential impact feature analysis module is used to perform potential physiological index impact correlation weight analysis on genomics data based on physiological index data to generate potential physiological index impact correlation weight data; and use the potential physiological index impact correlation weight data to perform individual physiological index weighting processing on the physiological index data to obtain individual physiological index data; The individual chronic disease influencing factor analysis module is used to analyze and process the pathological influencing factors of chronic disease types based on electronic health record data and individual physiological indicator data to generate individual chronic disease influencing factors; perform quantitative calculations on individual chronic disease influencing factors to generate quantitative data of individual chronic disease influencing factors; An individual chronic disease composite interactive risk factor analysis module is used to analyze and process the composite interactive risk factors of individual chronic disease influencing factors based on the quantitative data of individual chronic disease influencing factors, and generate individual chronic disease composite interactive risk factor data; The individual chronic disease risk assessment module is used to perform individual chronic disease risk assessment processing based on individual chronic disease composite interactive risk factor data to generate individual chronic disease risk assessment data.

[0005] Furthermore, the physiological index analysis module includes the following functions: Performing time series characteristic analysis of medical physiological monitoring signals according to the medical physiological monitoring signals to generate time series characteristic data of the medical physiological monitoring signals; Physiological index conversion processing is performed according to the time series characteristic data of the medical physiological monitoring signal to generate physiological index data.

[0006] Furthermore, the physiological index potential impact feature analysis module includes the following functions: Acquire genome reference data from a medical database, wherein the genome reference data includes genome sequence reference data and genome protein structure-function relationship data; Performing whole genome sequencing analysis on genomic data to extract coding region information of genomic data; using genome sequence reference data to annotate the coding region information of genomic data to generate gene coding region identification data; Performing genome coding region-specific low-dimensional mapping reconstruction processing on the gene coding region identification data to generate mapped genome coding region identification data; According to the physiological indicator data, the mapping genome coding region identification data is analyzed for the associated target mapping node of the physiological indicator to generate the target mapping genome coding region identification data; Based on the genome protein structure-function relationship data, the target mapping genome coding region identification data is analyzed for potential physiological index impact relationship, generating potential physiological index impact relationship data, and designing potential physiological index impact association weight data through the potential physiological index impact relationship data; The physiological indicator data are weighted by individual physiological indicator weights using the potential physiological indicator impact association weight data to obtain individual physiological indicator data.

[0007] Furthermore, the specific low-dimensional mapping reconstruction process of the genome coding region on the gene coding region identification data includes: Perform coding region specific analysis based on gene coding region identification data to generate coding region specific data; Perform coding region specific heterogeneous feature analysis based on coding region specific data to generate coding region specific heterogeneous feature data; Perform coding region-specific mapping relationship analysis based on coding region-specific heterogeneous feature data to generate coding region-specific mapping decisions; The genome coding region identification data is subjected to genome coding region-specific low-dimensional mapping reconstruction through coding region-specific mapping decision to generate mapped genome coding region identification data.

[0008] Furthermore, the analysis of the potential physiological index impact relationship of the target mapped genome coding region identification data based on the genome protein structure-function relationship data includes: Based on the preset convolutional neural network algorithm and genome protein structure-function relationship data, the mapping relationship between the protein structure and function of the genome is analyzed to obtain a genome function prediction model; Transmitting the target mapped genome coding region identification data to the genome function prediction model for target genome function prediction processing to generate target genome function prediction data; The potential physiological index impact relationship analysis is performed based on the target genome function prediction data to generate potential physiological index impact relationship data.

[0009] Furthermore, the individual chronic disease influencing factor analysis module includes the following functions: Performing historical pathology analysis on the electronic health record data to generate historical pathology analysis data, and performing historical pathology feature analysis based on the historical pathology analysis data to generate historical pathology feature data; Based on historical pathological characteristic data and individual physiological index data as individual pathological influencing factors, and using preset individual chronic disease composite interactive risk factor data to classify individual pathological influencing factors into pathological influencing factors of chronic disease types, individual chronic disease influencing factors are generated; The individual chronic disease impact factor evaluation algorithm is used to perform quantitative calculation and processing on the individual chronic disease impact factor to generate individual chronic disease impact factor quantitative data.

[0010] Furthermore, the individual chronic disease impact factor evaluation algorithm is as follows: ; ; ; ; In the formula, Expressed as the quantitative value of the individual chronic disease impact factor, It is expressed as the number of comprehensive influencing parameters of physiological indicators, Expressed as Comprehensive impact weight information of physiological indicators, Expressed as The impact value of chronic diseases on physiological indicators, Expressed as A time series sample set of chronic disease impact values ​​of physiological indicators, Expressed as The impact of chronic diseases on physiological indicators over time The standardized function of Expressed as the number of historical pathological characteristic parameters, Expressed as The weight information of historical pathological characteristic parameters, Expressed as The chronic disease impact value of the historical pathological characteristic parameters, Expressed as Historical pathological characteristic parameters, Expressed as The pathology-specific influence function of the historical pathological characteristic parameters, Expressed as The pathology-specific influence function adjustment coefficient of the historical pathological characteristic parameters, Represented as the chronic disease risk weight information of an individual’s genome, Expressed as the genomic impact value of chronic diseases, Chronic disease impact adjustment factors expressed as genomic transformations, Gene signature Individual physiological indicators Genetic and physiological factors of chronic diseases affect the conversion function.

[0011] Furthermore, the individual chronic disease composite interactive risk factor analysis module includes the following functions: Conduct correlation feature analysis on individual chronic disease influencing factors to generate correlation feature data on individual chronic disease assessment factors; Design a preliminary knowledge graph of individual chronic disease impact factors based on the quantitative data of individual chronic disease impact factors, and generate a preliminary individual chronic disease impact assessment graph; Based on the individual chronic disease assessment factor correlation characteristic data, the preliminary individual chronic disease impact assessment map is processed with correlation relationship identification to generate a correlated individual chronic disease impact assessment map; Analyze the related promoting factors of chronic diseases based on the related individual chronic disease impact assessment map to generate data on related promoting factors of chronic diseases; Using the data of chronic disease-related promoting influencing factors, the corresponding graph nodes in the individual chronic disease impact assessment graph are analyzed and processed for the nonlinear composite interactive characteristics of chronic disease influencing factors to generate composite interactive characteristic data of chronic disease influencing factors; According to the composite interactive characteristic data of chronic disease influencing factors, the quantitative data of individual chronic disease influencing factors are analyzed and processed into composite interactive risk factors of individual chronic disease influencing factors to generate individual chronic disease composite interactive risk factor data.

[0012] Furthermore, the individual chronic disease risk assessment module includes the following functions: Monitor and process the data of composite interactive risk factors of individual chronic diseases to generate monitoring data of composite interactive risk factors of individual chronic diseases; Perform cluster analysis of chronic disease risk factors on monitoring data of individual chronic disease composite interactive risk factors to generate cluster data of individual chronic disease composite interactive risk factors; Perform interactive risk factor cluster outlier trend slope analysis based on individual chronic disease composite interactive risk factor cluster data to generate interactive risk factor cluster outlier trend slope data; Based on the preset chronic disease risk trend level threshold, the interactive risk factor cluster outlier trend slope data is processed by trend slope classification to generate interactive risk factor cluster outlier trend slope classification data; Individual chronic disease risk assessment is performed based on individual chronic disease composite interactive risk factor data and interactive risk factor cluster outlier trend slope grading data to generate individual chronic disease risk assessment data.

[0013] Furthermore, the performing interactive risk factor clustering outlier trend slope analysis based on individual chronic disease composite interactive risk factor clustering data includes: Performing interactive risk factor cluster outlier analysis based on individual chronic disease composite interactive risk factor cluster data to generate interactive risk factor cluster outlier data; Perform cluster outlier trend analysis on the interactive risk factor cluster outlier data to generate interactive risk factor cluster outlier trend data; The cluster outlier trend slope analysis is performed on the interactive risk factor cluster outlier trend data to generate the interactive risk factor cluster outlier trend slope data.

[0014] The present invention provides intelligent and comprehensive support for the prevention, diagnosis and personalized health management of chronic diseases through the synergy of various modules in the chronic disease risk assessment system. The multi-source physiological sign data acquisition module integrates the data of the medical cloud service platform, including electronic health record data, medical physiological monitoring signals and genomics data, to achieve comprehensive monitoring of the patient's physiological state, and provide multi-dimensional and comprehensive basic data support for subsequent indicator analysis. The time series feature analysis of the medical physiological monitoring signal in the physiological indicator analysis module can deeply explore the time change pattern and dynamic characteristics in the signal, reveal the subtle changes in the individual's health status, and effectively capture the change trend and fluctuation law of physiological parameters such as heart rate, blood pressure, and respiratory rate by performing time series feature analysis on the medical physiological monitoring signal, thereby providing a reliable basis for early warning and accurate diagnosis of chronic diseases. Based on the time series feature data, the physiological indicator conversion processing can convert complex physiological monitoring signals into intuitive and quantifiable physiological indicator data, which enhances the operability of the data and the accuracy of the analysis. The physiological indicator potential impact feature analysis module provides reliable reference information for the in-depth analysis of genomic data by obtaining the genome reference data of the medical database. By introducing genome sequence reference data and genome protein structure-function relationship data, the gene characteristics and functional associations related to chronic diseases can be accurately identified. The genome coding region information is extracted through whole genome sequencing analysis, and the genome sequence reference data is used to annotate the gene coding region, effectively identifying the key regions in the genome, quickly locking the gene regions closely related to the pathogenesis of chronic diseases, and improving the pertinence and efficiency of data analysis. The gene coding region identification data is processed by specific low-dimensional mapping reconstruction of the genome coding region, and the key features related to physiological indicators are extracted from the high-dimensional complex genome data, reducing the computational complexity while retaining the specific gene features that are highly related to chronic disease risk assessment. The mapping genome coding region identification data is analyzed by associating target mapping nodes using physiological indicator data to clarify the potential connection between genome data and physiological indicators. The generated target mapping genome coding region identification data provides structured input for subsequent potential impact relationship analysis. Based on the data of genomic protein structure-function relationship, the target mapping genome coding region identification data is analyzed for potential physiological index influence relationship, revealing the path and mechanism of gene action on physiological indexes. By generating potential physiological index influence relationship data and designing association weights, the correlation between genes and physiological indexes is quantified, providing accurate biological explanation and weight distribution basis for multidimensional risk assessment of chronic diseases. The physiological index data is weighted by individual physiological index weights using the potential physiological index influence association weight data, realizing individualized adjustment of physiological index data, transforming the group model into an individualized assessment tool, and improving the accuracy and applicability of the system in chronic disease risk assessment.The individual chronic disease influencing factor analysis module can extract key information from the patient's medical records by performing historical pathological analysis on the electronic health record data, and reveal the historical pathological characteristics related to chronic diseases through historical pathological feature analysis, combining the patient's past health status with chronic disease risk assessment, and providing an important basis for identifying pathological patterns and trends. Based on historical pathological feature data and individual physiological index data, it is used as individual pathological influencing factors, and combined with the preset individual chronic disease composite interactive risk factor data to classify the pathological influencing factors of chronic disease types, which can clarify the key influencing factors of different chronic disease types, reveal the differences in influencing factors between different chronic diseases, enhance the pertinence and accuracy of the evaluation results, and provide guidance for personalized intervention. The individual chronic disease influencing factor evaluation algorithm is used to quantify the individual chronic disease influencing factors. This quantification process converts multidimensional pathological feature data into intuitive evaluation indicators, which intuitively reflects the physiological indicators related to different chronic diseases and the level of pathology on individual chronic disease risk. In the individual chronic disease composite interactive risk factor analysis module, the potential relationship between the influencing factors can be revealed by analyzing the correlation characteristics of the individual chronic disease influencing factors. By mining the correlation characteristics, the interaction of multiple factors on chronic disease risks can be more comprehensively identified. According to the quantitative data of individual chronic disease influencing factors, the preliminary knowledge graph is designed to effectively construct the logical relationship and association path between chronic disease influencing factors, providing a structured framework for individual chronic disease risk assessment. Based on the correlation feature data of individual chronic disease assessment factors, the preliminary knowledge graph is processed for association relationship identification, and the associated individual chronic disease impact assessment graph is generated. The structure and information of the graph are further optimized, the dynamic relationship between chronic disease influencing factors is accurately identified, the logical clarity and practicality of the knowledge graph are improved, and a solid foundation for association analysis is laid. According to the associated individual chronic disease impact assessment graph, the chronic disease associated promoting factors are analyzed to identify the key factors that promote chronic disease risks, clarify the risk promoting factors, provide targeted guidance for chronic disease prevention and control, and optimize the key analysis indicators in subsequent evaluations. The nonlinear composite interaction feature analysis of the corresponding nodes in the evaluation graph is performed using the data of chronic disease associated promoting influencing factors. The complex multi-factor interaction pattern is captured through nonlinear analysis methods, which makes up for the shortcomings of linear analysis of the interaction between individual chronic disease influencing factors and improves the accuracy and comprehensiveness of the evaluation. According to the composite interaction feature data, the composite interaction risk factor analysis of the quantitative data of individual chronic disease influencing factors is performed to fully integrate complex multidimensional data. The individual chronic disease risk assessment module monitors and processes the data of individual chronic disease composite interactive risk factors, and captures the fluctuation trend of chronic disease-related risk factors in real time, providing a basic guarantee for early detection of potential problems. Cluster analysis of chronic disease risk factors on monitoring data, summarizing and classifying different risk factors, helps to identify common laws and unique patterns, and improves the comprehensiveness and accuracy of risk assessment.Based on cluster data, we conduct trend slope analysis of cluster outliers of interactive risk factors to capture the trend changes of abnormal risk factors and provide key inputs for the early warning system. Based on the preset chronic disease risk trend level threshold, we perform trend slope classification on trend slope data, clarify the risk level through classification strategy, and provide hierarchical information for individual chronic disease risk assessment, so as to facilitate the formulation of higher priority intervention measures. Finally, we conduct individual chronic disease risk assessment based on individual chronic disease composite interactive risk factor data and trend slope classification data. By combining multidimensional data and classification results, we can accurately assess the severity of chronic diseases.

[0015] The beneficial effect of the present application is that the chronic disease risk assessment system of the present invention uses multi-source physiological sign data, including electronic health records, physiological signals, and genomes, to conduct collaborative assessment of chronic disease risks to reflect the health status of individuals. In addition, the interaction of multiple factors such as physiology and abnormalities can be used to explore the pathological mechanisms and risk characteristics of chronic disease trends, so as to timely discover and intervene in the early symptoms of chronic diseases, provide important support for individual chronic disease health management and accurate assessment, and also have important significance for reducing the medical burden and improving the level of chronic disease health management. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A schematic diagram of the module flow of a chronic disease risk assessment system of the present invention; Figure 2 for Figure 1 Detailed functional flow chart of the potential impact feature analysis module of physiological indicators; Figure 3 for Figure 1 Detailed functional flow chart of the individual chronic disease complex interactive risk factor analysis module. DETAILED DESCRIPTION

[0017] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0018] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0019] It should be understood that although the terms "first", "second", etc. are used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly a second unit may be referred to as a first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0020] To achieve this, please refer to Figures 1 to 3 The present invention provides a chronic disease risk assessment system, comprising the following modules: A multi-source physiological sign data collection module is used to collect multi-source physiological sign data of patients through a medical cloud service platform to generate multi-source physiological sign data, wherein the multi-source physiological sign data includes electronic health record data, medical physiological monitoring signals and genomics data; A physiological index analysis module is used to analyze physiological indexes according to medical physiological monitoring signals and generate physiological index data; The physiological index potential impact feature analysis module is used to perform potential physiological index impact correlation weight analysis on genomics data based on physiological index data to generate potential physiological index impact correlation weight data; and use the potential physiological index impact correlation weight data to perform individual physiological index weighting processing on the physiological index data to obtain individual physiological index data; The individual chronic disease influencing factor analysis module is used to analyze and process the pathological influencing factors of chronic disease types based on electronic health record data and individual physiological indicator data to generate individual chronic disease influencing factors; perform quantitative calculations on individual chronic disease influencing factors to generate quantitative data of individual chronic disease influencing factors; An individual chronic disease composite interactive risk factor analysis module is used to analyze and process the composite interactive risk factors of individual chronic disease influencing factors based on the quantitative data of individual chronic disease influencing factors, and generate individual chronic disease composite interactive risk factor data; The individual chronic disease risk assessment module is used to perform individual chronic disease risk assessment processing based on individual chronic disease composite interactive risk factor data to generate individual chronic disease risk assessment data.

[0021] In the embodiment of the present invention, reference Figure 1 The above is a schematic diagram of the module flow of a chronic disease risk assessment system of the present invention. In this embodiment, the chronic disease risk assessment system includes: S1: a multi-source physiological sign data collection module, used to collect multi-source physiological sign data of patients through a medical cloud service platform to generate multi-source physiological sign data, wherein the multi-source physiological sign data includes electronic health record data, medical physiological monitoring signals and genomics data; In an embodiment of the present invention, a medical cloud service platform is used to collect multi-source physiological sign data stored by patients in medical communities and other places, including the patient's electronic health record (EHR), medical physiological monitoring signals, and genomic data. Specifically, the electronic health record data covers the patient's medical history, medical records, drug use history, laboratory test results, imaging data and other information, providing a comprehensive health profile of the patient. Medical physiological monitoring signals collect the patient's physiological indicators in real time through various sensors (such as electrocardiogram (ECG), blood oxygen saturation (SpO2), blood pressure monitoring, body temperature, etc.) to reflect their current physiological state and health status. Genomic data comes from genome sequencing and contains the patient's genetic information, which is mainly used to analyze genetic risk factors and gene mutations or expression levels related to chronic diseases. All these data are aggregated and stored and processed by the medical cloud service platform to generate unified multi-source physiological sign data, providing basic data support for subsequent chronic disease risk assessment and personalized medical decision-making. This process ensures efficient data collection and synchronization, and provides comprehensive and accurate physiological sign information for subsequent analysis and evaluation of the system.

[0022] S2: Physiological index analysis module, used to analyze physiological indexes according to medical physiological monitoring signals and generate physiological index data; In an embodiment of the present invention, a time series feature analysis of a medical physiological monitoring signal is performed based on a medical physiological monitoring signal. The analyzed medical physiological monitoring signal is a short-term medical monitoring signal, that is, a medical physiological monitoring signal within a cycle or a day. The medical physiological monitoring signal includes an electrocardiogram (ECG) signal, a blood pressure signal, and a respiratory rate signal. A continuous ECG signal is obtained, and data is recorded at a sampling rate of 500 Hz. The ECG time series feature is mapped to a heart rate variability (HRV) index through a machine learning model (such as a support vector machine SVM). The PWTT data is used to estimate the arteriosclerosis index, and the respiratory signal data is used to calculate the vital capacity and ventilation efficiency, and the heart rate variability, arteriosclerosis index, vital capacity and ventilation efficiency, etc. are obtained, and finally the physiological index data is formed.

[0023] S3: A physiological indicator potential impact feature analysis module, which is used to perform potential physiological indicator impact association weight analysis on genomics data based on physiological indicator data to generate potential physiological indicator impact association weight data; and to perform individual physiological indicator weighting processing on physiological indicator data using potential physiological indicator impact association weight data to obtain individual physiological indicator data; In an embodiment of the present invention, the genome reference data of the medical database is obtained, and the latest human genome reference data set GRCh38 is downloaded by accessing the medical database. The data set provides complete chromosome sequence information, including nucleotide sequence and chromosome position markers, to ensure consistency with whole genome sequencing. The database retrieval function is used to extract protein structure-function relationship data, including protein sequence, protein three-dimensional structure model, functional annotation and functional path data associated with gene variation from the UniProt database. The gene coding region is taken as the key area, all variations in the coding region are screened out, and the exon number and variation effect (such as missense mutation or nonsense mutation) are recorded. The gene coding region identification data is subjected to a specific low-dimensional mapping reconstruction process of the genome coding region, and the UMAP algorithm is used for dimensionality reduction processing. The input data includes high-dimensional features of each variation, such as chromosome position, variation type, exon number, amino acid change, functional effect, etc., in a matrix format (each row is a variation, each column is a feature). By performing low-dimensional manifold mapping on the variation features in the high-dimensional space, coordinate points in the two-dimensional space are generated, the specificity and high correlation features of the genome coding region are retained, and the mapping genome coding region identification data is generated. According to the physiological index data, the mapping genome coding region identification data is analyzed for the associated target mapping nodes of the physiological indexes. The key physiological index data such as blood pressure, blood sugar, heart rate, etc. of the patient are used as input parameters. The previously generated genome coding region two-dimensional mapping identification data is analyzed for the association of the target nodes by using the association characteristics with the genome coding region identification data. The target mapping genome coding region identification data is used for joint analysis with the genome protein function relationship data. Protein function pathways, protein interaction network information, and key functional region data associated with chronic diseases are extracted. For example, for the BRCA1 gene, the DNA repair function pathway of its encoded protein and the effect of its mutation on the downstream signaling pathway are analyzed. The centrality parameters (such as Degree and Betweenness) of each node are calculated using the protein interaction network provided by the STRING database. Combined with the protein-physiological index association model, the potential influence of each gene node on the change of physiological indicators is quantified, and the results are stored as a potential influence relationship data table, which includes the gene name, protein function description, and the name of the associated physiological index. The protein function description is analyzed by logistic regression, converted into the parameter of the protein function on the physiological index, and the potential physiological index influence association weight design of the protein function is combined with the analysis results and the correlation coefficient. The physiological indicator data are weighted by individual physiological indicator weights using the potential physiological indicator impact association weight data to obtain individual physiological indicator data.

[0024] S4: Individual chronic disease influencing factor analysis module, used to analyze and process the pathological influencing factors of chronic disease types based on electronic health record data and individual physiological index data, and generate individual chronic disease influencing factors; perform quantitative calculation and processing on individual chronic disease influencing factors, and generate quantitative data of individual chronic disease influencing factors; In an embodiment of the present invention, in an embodiment of the present invention, the electronic health record (EHR) data is subjected to historical pathological analysis processing, and a parsing algorithm based on natural language processing (NLP) is used to automatically extract information such as disease type, symptoms, treatment plan, etc. from the medical record text, and the corresponding time nodes and treatment results are marked. After the historical pathological analysis data is generated, a historical pathological feature analysis is performed, including pattern recognition and feature extraction of the parsed historical pathological data to reveal key pathological features related to chronic diseases, and obtain historical pathological feature data. The historical pathological feature data is combined with individual physiological index data as individual pathological influencing factors. These data include but are not limited to the patient's medical history information, current physiological indicators (such as weight, blood sugar, blood lipids, etc.) and their historical changes. Combining the historical pathological feature data and the physiological index data, a series of pathological influencing factors are defined. The individual chronic disease influencing factors are quantitatively calculated using an individual chronic disease influencing factor evaluation algorithm. First, based on the classification of pathological influencing factors, the evaluation algorithm assigns a weight to each influencing factor, which reflects the contribution of the factor to the occurrence of chronic diseases. The setting of the weight can be determined by a large amount of clinical data analysis and statistical models, and is usually estimated using methods such as regression analysis, machine learning, or expert systems. For each influencing factor, such as high blood sugar and obesity, the evaluation algorithm quantifies its influence through a series of statistical indicators (such as risk ratio, standardized regression coefficient, etc.). Each pathological influencing factor will be assigned a quantitative value. After completing the quantitative calculation, the algorithm will integrate these quantitative data into a set of quantitative data of individual chronic disease influencing factors. This data can be used as the basis for subsequent risk assessment and decision support to help doctors determine the chronic disease risk level of patients.

[0025] S5: Individual chronic disease composite interactive risk factor analysis module, used to analyze and process the composite interactive risk factors of individual chronic disease influencing factors based on the quantitative data of individual chronic disease influencing factors, and generate individual chronic disease composite interactive risk factor data; In an embodiment of the present invention, an association feature analysis is performed on individual chronic disease influencing factors to identify which influencing factors have a strong statistical correlation with each other. For example, in the assessment of diabetes, it is found that blood sugar levels are strongly correlated with insulin resistance, and there is also a significant correlation between hypertension and obesity. The Pearson correlation coefficient is used to calculate the linear relationship between different factors, and these factors are grouped into categories, such as performing association degree grouping, and further analyzing the degree of correlation between each group of factors. In this way, the association feature data of individual chronic disease assessment factors are obtained, and the preliminary knowledge map of individual chronic disease influencing factors is designed based on the quantitative data of individual chronic disease influencing factors. Using the graph theory model, the individual chronic disease influencing factors are used as nodes in the graph, and the edges between the nodes represent the relationship between different influencing factors. The weight of the edge is determined by the correlation data or statistical analysis results obtained in the association feature analysis, and an individual chronic disease impact assessment map is constructed, and the relationship between chronic disease influencing factors is visualized in the form of a map. Based on the correlation feature data of individual chronic disease assessment factors, the preliminary individual chronic disease impact assessment map is processed with correlation relationship identification, and the relationship in the map is adjusted using statistical methods or machine learning algorithms (such as Bayesian networks, decision trees, etc.) to ensure that the correlation relationship between each node reflects the strength of the causal effect. The generated associated individual chronic disease impact assessment map more accurately presents the strength of the correlation between each factor. According to the associated individual chronic disease impact assessment map, the associated promoting factors of chronic diseases are analyzed, and the factors of each map are analyzed to promote the risk of chronic diseases in patients, and are marked as promoting factors of chronic diseases. By analyzing the weighted influence relationship between nodes, the key factors that promote pathological changes are extracted. Based on these factors, chronic disease associated promoting factor data are generated. Using the chronic disease associated promoting influencing factor data, the potential interactions between each node in the map (i.e., chronic disease influencing factors) are identified. These influencing factors include but are not limited to body mass index (BMI), blood sugar level, blood pressure value, cholesterol, etc., and these data are analyzed by statistical models or machine learning algorithms. By inputting individual chronic disease influencing factor data into the trained nonlinear model, the interaction characteristics between each pair of factors are calculated and extracted, including the interaction intensity between each chronic disease influencing factor and their compound effects under nonlinear modeling, so as to obtain the composite interaction characteristics of each individual chronic disease assessment factor. Combined with these composite interaction feature data, the node relationship in the graph is updated to reflect the more sophisticated and complex interactions between chronic disease influencing factors.According to the composite interactive characteristic data of chronic disease influencing factors, the composite interactive risk factor analysis of individual chronic disease influencing factors is performed on the quantitative data of individual chronic disease influencing factors, and the quantitative data of individual chronic disease influencing factors are collected, including various physiological indicators and medical history information related to chronic diseases, such as individual body mass index (BMI), blood sugar level, blood pressure value, cholesterol content, exercise habits, eating habits and family medical history. After data collection, these factors will be converted into quantitative values ​​to identify the chronic disease risks of specific combinations and generate risk scores for each factor combination. Based on the analysis results, individual chronic disease composite interactive risk factor data are generated. These data reflect the overall risk level of individuals under the interaction of multiple chronic disease risk factors. The data are usually presented in the form of risk scores or weight values, reflecting the risks of individuals under different factor combinations.

[0026] S6: Individual chronic disease risk assessment module, used to perform individual chronic disease risk assessment processing based on individual chronic disease composite interactive risk factor data to generate individual chronic disease risk assessment data.

[0027] In an embodiment of the present invention, the individual chronic disease composite interactive risk factor data generated in the above steps are continuously monitored, and the monitoring results are saved in a database to obtain individual chronic disease composite interactive risk factor monitoring data. After completing the data monitoring, cluster analysis of chronic disease risk factors is performed next. Chronic disease risk factors are classified and processed by clustering algorithms (such as K-means, hierarchical clustering, etc.). The goal is to cluster and group individual monitoring data. Taking blood sugar and BMI as examples, blood sugar and BMI related data are clustered and grouped, and outliers of each cluster group are monitored. And according to the changing trend of these outliers, the trend slope analysis of the clustering outliers of the interactive risk factor is performed to evaluate whether the risk factor is changing, and whether the trend of the change is upward or downward. For example, in the individual chronic disease composite interactive risk factor data within 6 months of continuous monitoring, if the blood sugar level and BMI of the patient individual in a certain cluster continue to increase, and this change trend is significant, then the individual's blood sugar and BMI changes will be considered as outliers, and the trend slope of the patient will be calculated. The trend slope can be calculated by linear regression or other correlation analysis methods. If the trend slope of the individual is positive and greater than 0.1 in the overall monitoring interval or each monitoring node, it means that the risk of chronic diseases is increasing. The trend slope of the outlier is graded by comparing with the preset chronic disease risk trend level threshold. These thresholds are set according to the clinical standards of chronic disease risk. The purpose is to divide the risk changes into different levels, classify each monitored interactive risk factor into different risk levels according to its change rate, and then determine the patient's chronic disease risk status. Considering the individual's chronic disease composite interactive risk factor data and trend slope classification data, the patient's overall chronic disease risk is evaluated. According to the trend slope classification data of the outlier, the change trend of each risk factor is combined with its corresponding chronic disease influencing factor, and a comprehensive evaluation is performed based on the change trend of these indicators and the patient's historical pathological data, and finally the individual chronic disease risk assessment data is obtained.

[0028] Furthermore, the physiological index analysis module includes the following functions: Performing time series characteristic analysis of medical physiological monitoring signals according to the medical physiological monitoring signals to generate time series characteristic data of the medical physiological monitoring signals; Physiological index conversion processing is performed according to the time series characteristic data of the medical physiological monitoring signal to generate physiological index data.

[0029] In an embodiment of the present invention, a time series feature analysis of a medical physiological monitoring signal is performed based on a medical physiological monitoring signal. The analyzed medical physiological monitoring signal is a short-term medical monitoring signal, that is, a medical physiological monitoring signal within a cycle or a day. The medical physiological monitoring signal includes an electrocardiogram (ECG) signal, a blood pressure signal, and a respiratory rate signal. A continuous ECG signal is obtained and data is recorded at a sampling rate of 500 Hz. High-frequency noise and baseline drift are eliminated by filtering technology, such as wavelet transform denoising. Key time series features, such as the amplitude of the QRS complex and the RR interval, are extracted from the ECG signal; the pulse wave transmission time (PWTT) is calculated for the blood pressure signal; and the respiratory cycle and amplitude changes are analyzed for the respiratory signal. The extracted features are integrated to form a time series feature data set, which is structured and stored in a database for subsequent analysis. The ECG time series features are mapped to heart rate variability (HRV) indicators through a machine learning model (such as a support vector machine SVM). The PWTT data is used to estimate the arteriosclerosis index, and the respiratory signal data is used to calculate the vital capacity and ventilation efficiency, thereby obtaining the heart rate variability, arteriosclerosis index, vital capacity and ventilation efficiency, and finally forming the physiological indicator data.

[0030] Furthermore, the physiological index potential impact feature analysis module includes the following functions: Acquire genome reference data from a medical database, wherein the genome reference data includes genome sequence reference data and genome protein structure-function relationship data; Performing whole genome sequencing analysis on genomic data to extract coding region information of genomic data; using genome sequence reference data to annotate the coding region information of genomic data to generate gene coding region identification data; Performing genome coding region-specific low-dimensional mapping reconstruction processing on the gene coding region identification data to generate mapped genome coding region identification data; According to the physiological indicator data, the mapping genome coding region identification data is analyzed for the associated target mapping node of the physiological indicator to generate the target mapping genome coding region identification data; Based on the genome protein structure-function relationship data, the target mapping genome coding region identification data is analyzed for potential physiological index impact relationship, generating potential physiological index impact relationship data, and designing potential physiological index impact association weight data through the potential physiological index impact relationship data; The physiological indicator data are weighted by individual physiological indicator weights using the potential physiological indicator impact association weight data to obtain individual physiological indicator data.

[0031] As an example of the present invention, refer to Figure 2 As shown, Figure 1The function flow diagram of the physiological index analysis module in this example includes: S301: Acquire genome reference data from a medical database, wherein the genome reference data includes genome sequence reference data and genome protein structure-function relationship data; In the embodiment of the present invention, the latest human genome reference data set GRCh38 is downloaded by accessing the medical database. The data set provides complete chromosome sequence information, including nucleotide sequence and chromosome position markers, to ensure consistency with whole genome sequencing. The database retrieval function is used to extract protein structure-function relationship data, including protein sequences, protein three-dimensional structure models, functional annotations, and functional pathway data associated with gene mutations from the UniProt database. The genome sequence reference data is stored in FASTA format, and the protein structure function data is stored in TSV format, and all files are ensured to meet bioinformatics standards for subsequent calculation and analysis.

[0032] S302: performing whole genome sequencing analysis on the genomics data to extract coding region information of the genomics data; using the genome sequence reference data to annotate the coding region information of the genomics data to generate gene coding region identification data; In an embodiment of the present invention, based on the second-generation sequencing technology (such as Illumina HiSeq 4000), DNA is extracted from the patient's peripheral blood sample, and the DNA is fragmented into 200-300 bp fragments by enzyme digestion and library construction, and a sequencing library is generated by PCR amplification. The short-read sequence data obtained by the sequencer is aligned to the GRCh38 reference genome by BWA software to generate an alignment file (BAM format). The GATK toolbox is used to perform quality control, duplicate sequence removal and alignment optimization to ensure that the generated variant detection file (VCF format) contains SNP and small Indel information. For each variant site, the chromosome number, position, base variant type and support depth (Reads Depth) are recorded to ensure that the extracted variant data is accurate. The generated VCF file is annotated by the ANNOVAR tool, and the GRCh38 gene coordinates and exon definition file (GTF format) are used to functionally classify the variants. Focusing on the gene coding region, all mutations in the coding region are screened out, and the exon number and mutation impact (such as missense mutation or nonsense mutation) are recorded. The generated gene coding region identification data is stored in a format that includes the gene name, mutation location, exon number, amino acid change type, and affected function.

[0033] S303: performing genome coding region-specific low-dimensional mapping reconstruction processing on the gene coding region identification data to generate mapped genome coding region identification data; In an embodiment of the present invention, based on the generated gene coding region identification data, the UMAP algorithm is used for dimensionality reduction processing. The input data includes high-dimensional features of each variation, such as chromosome position, variation type, exon number, amino acid change, functional impact, etc., in a matrix format (each row is a variation, each column is a feature). By mapping the variation features in the high-dimensional space to a low-dimensional manifold, coordinate points in a two-dimensional space are generated to retain the specificity and high correlation features of the genome coding region. For example: a gene variation record "BRCA1.5266dupC" in the input data is mapped to a two-dimensional space coordinate (x=12.345, y=9.876) and stored as a two-dimensional point set (CSV format) for subsequent physiological index association analysis. When reducing the dimension, set the parameters n_neighbors=15 and min_dist=0.1 to ensure that the local neighborhood structure is complete. The two-dimensional identification data generated by the mapping includes a set of coordinate points for all genomic variations, which is convenient for visualization and further calculation.

[0034] S304: performing a target mapping node analysis of the physiological indicator-related mapping genome coding region identification data according to the physiological indicator data to generate target mapping genome coding region identification data; In an embodiment of the present invention, the patient's key physiological index data such as blood pressure, blood sugar, heart rate, etc. are used as input parameters, and the previously generated genome coding region two-dimensional mapping identification data is used to perform association analysis of the target node by utilizing its association characteristics with the genome coding region identification data. The correlation score of each physiological index and each node in the genome identification data is calculated by Pearson correlation analysis, and the correlation threshold is set to 0.7, and nodes with significant positive correlation are retained. For example, the genome identification data node "BRCA1.5266dupC" significantly associated with blood pressure data is assigned a correlation coefficient of 0.85. The collection of target mapping nodes forms a new genome identification data file, which contains all gene nodes significantly associated with physiological indicators and their correlation coefficients. The data is stored in the form of a relational table, recording the gene name, node number, physiological index category and correlation coefficient.

[0035] S305: performing potential physiological indicator influence relationship analysis on the target mapping genome coding region identification data based on the genome protein structure-function relationship data, generating potential physiological indicator influence relationship data, and designing potential physiological indicator influence association weight data through the potential physiological indicator influence relationship data; In an embodiment of the present invention, the target mapping genome coding region identification data is used to perform a joint analysis with the genome protein function relationship data. Protein function pathways, protein interaction network information, and key functional area data associated with chronic diseases are extracted. For example, for the BRCA1 gene, the DNA repair function pathway of its encoded protein and the effect of its mutation on the downstream signaling pathway are analyzed. The centrality parameters (such as Degree and Betweenness) of each node are calculated using the protein interaction network provided by the STRING database. Combined with the protein-physiological indicator association model, the potential impact of each gene node on the change of physiological indicators is quantified, and the results are stored as a potential impact relationship data table, the format of which includes gene name, protein function description, and associated physiological indicator name. The protein function description is analyzed by logistic regression, converted into the parameter of the protein function on the physiological indicator, and the potential physiological indicator impact association weight design of the protein function is combined with the analysis results and the correlation coefficient.

[0036] S306: Using the potential physiological indicator impact association weight data, individual physiological indicator weight assignment processing is performed on the physiological indicator data to obtain individual physiological indicator data.

[0037] In an embodiment of the present invention, based on the potential physiological index impact weight data, the individualized weight of each indicator is calculated one by one for the patient's physiological index data. For example, using the weight of 0.85 for blood pressure and BRCA1 gene variation and the weight of 0.78 for protein function, the two are weighted and calculated to generate an individualized weight value of 0.83. After the weight data of all physiological indicators are normalized, the original physiological index value is corrected according to the weighted average formula. For example, the patient's original blood pressure is 140 mmHg, and the impact on chronic diseases related to blood pressure is U value. After weight correction, it becomes 0.83 multiplied by U value for the impact on chronic diseases related to blood pressure. The generated individual physiological index data records the numerical comparison and weight value before and after the specific weight adjustment, which is used for subsequent chronic disease risk assessment. The data is stored as a structured table, including patient ID, physiological index name, original value, weight correction value and final correction value.

[0038] Furthermore, the specific low-dimensional mapping reconstruction process of the genome coding region on the gene coding region identification data includes: Perform coding region specific analysis based on gene coding region identification data to generate coding region specific data; Perform coding region specific heterogeneous feature analysis based on coding region specific data to generate coding region specific heterogeneous feature data; Perform coding region-specific mapping relationship analysis based on coding region-specific heterogeneous feature data to generate coding region-specific mapping decisions; The genome coding region identification data is subjected to genome coding region-specific low-dimensional mapping reconstruction through coding region-specific mapping decision to generate mapped genome coding region identification data.

[0039] In an embodiment of the present invention, the sequence information and annotation data of all relevant genes are extracted according to the obtained gene coding region identification data. For each gene, especially the mutation site in the coding region, a detailed analysis is performed. For example, for the coding region of the gene BRCA1, its specific mutation site, such as c.5074G>A, is searched in the genome, and the variation frequency of the site is analyzed by bioinformatics methods. The frequency information of the variation can be obtained through the dbSNP database. If the variation has a high frequency (such as 5%) in a specific population, it is considered that the variation has a strong specificity. The specificity of the site is analyzed by the conservative score of the gene. For example, the conservatism score of the site is calculated using the PhyloP or GERP scoring tool. If its value is higher than 0.9, it means that the site has a high degree of conservation in species evolution, and it is further considered that it is more functional in the organism and has a greater potential impact on chronic diseases. On the basis of the specific data of the gene coding region, the coding region specific heterogeneous feature analysis is further performed to identify the structural or functional features associated with the gene variation. Based on the generated specific data, the functional characteristics of the specific region are analyzed by the three-dimensional structural information of the gene. For example, if the c.5074G>A site of the BRCA1 gene causes an amino acid change in the encoded protein (such as from glutamic acid to alanine), it is necessary to simulate the effect of the mutation on the three-dimensional structure of the protein through molecular modeling software such as PyMOL, and further calculate the effect of the mutation on the stability of the protein structure (such as protein folding ability, hydrogen bond formation, etc.) to evaluate its functional changes. For example, the analysis of the changes in the tertiary structure of the protein caused by its nucleotide substitution found that the mutation caused a 15% decrease in structural stability. Based on these parameters, a heterogeneous feature matrix is ​​established to classify and summarize the sequence characteristics and functional changes, and the first two principal components are extracted as heterogeneous features. The heterogeneous features of the mutation site are analyzed, and the variation of different patients at this site is stored in the GnomAD database for statistics to identify whether the variation has a higher frequency in a specific population. In addition, it is also necessary to combine the analysis of whether the mutation is associated with a specific disease (such as breast cancer, cardiovascular disease, etc.), and calculate the heterogeneous feature score for each coding region-specific data. For example, the nucleotide composition, sequence mutation pattern, and structural and functional characteristics of each specific region are quantitatively analyzed. These scores include but are not limited to structural change scores, functional change scores, disease associations, etc., and are quantitatively analyzed in combination with information such as gene expression levels, mutation effects, and disease risks, ultimately generating coding region-specific heterogeneous feature data. After obtaining coding region-specific heterogeneous feature data, coding region-specific mapping relationship analysis is performed. The core goal is to generate a mapping decision model based on known gene mutation data, heterogeneous features, and their functional effects to predict which gene variants have potential effects on chronic diseases.Based on the collected heterogeneous feature data (such as structural change scores, functional change scores, disease association, etc.), the supervised learning algorithm is used to evaluate the impact of each feature on the coding region of the genome. For example, gene variants with a disease association score greater than 0.8 are set as high-risk areas, and those below 0.3 are set as low-risk areas. By training the supervised learning algorithm, a mapping relationship model is constructed, which can map gene coding regions based on specific gene mutations and their related functional characteristics (such as conservation, amino acid changes, etc.). For example, analysis shows that certain mutation sites (such as BRCA1 c.5074G>A) are closely related to the risk of breast cancer. At the same time, the structural change score of this site is 0.85, the functional change score is 0.78, and the final calculated disease association is 0.92. These data are mapped into the decision model, and the gene mutation is classified as a high-risk area based on the specificity of the genomic region. Based on this analysis method, the generated coding region-specific mapping decision can assign a mapping category to each gene variant, such as low risk, medium risk, and high risk, and evaluate the risk of different disease types by adjusting the corresponding thresholds. Based on the coding region-specific mapping decision, the genome coding region identification data is reconstructed by specific low-dimensional mapping to obtain the genome coding region identification data, which includes detailed information of each gene variation (such as gene position, mutation type, variation frequency, etc.). Then, based on the coding region-specific mapping decision model, those gene variations that belong to the high-risk category are selected and specifically mapped. For example, for the BRCA1 c.5074G>A variation marked as high-risk, it is mapped to a low-dimensional space through low-dimensional mapping techniques, such as principal component analysis, and mapping decisions. For example, relevant feature data (such as structural changes, functional changes, disease association, etc.) are selected based on high-risk variations, and then these feature data are mapped to two-dimensional or three-dimensional space. In this low-dimensional space, different gene variations will be classified, and similar variations will be assigned to the same region. Through this mapping process, the specific information of the genome coding region will be simplified into low-dimensional data, making the risk assessment of chronic diseases more intuitive and easy to understand. The mapped data will be used for further chronic disease risk assessment processing.

[0040] Furthermore, the analysis of the potential physiological index impact relationship of the target mapped genome coding region identification data based on the genome protein structure-function relationship data includes: Based on the preset convolutional neural network algorithm and genome protein structure-function relationship data, the mapping relationship between the protein structure and function of the genome is analyzed to obtain a genome function prediction model; Transmitting the target mapped genome coding region identification data to the genome function prediction model for target genome function prediction processing to generate target genome function prediction data; The potential physiological index impact relationship analysis is performed based on the target genome function prediction data to generate potential physiological index impact relationship data.

[0041] In an embodiment of the present invention, the mapping relationship between protein structure and function is analyzed based on the convolutional neural network (CNN) algorithm through the genome protein structure function relationship data to obtain a genome function prediction model. Convolutional neural network is selected as an analysis tool because it performs well in image recognition and high-dimensional data analysis, and is particularly suitable for processing protein structure data. The collected genome protein structure function relationship data includes information such as the three-dimensional structure, functional region, amino acid sequence, physicochemical properties of amino acids, mutation sites, etc. of the protein. In order to construct a convolutional neural network model, it is necessary to process these structural data and convert the three-dimensional structure of the protein into a data format suitable for input into the convolutional neural network. A common practice is to convert the protein structure into an image or matrix, for example, by extracting the structural information of the protein using a PDB (ProteinData Bank) format file and converting it into a matrix or image. The convolutional neural network learns and extracts features from these data through structures such as stacked convolutional layers and pooling layers, and the model training maps the complex functions of proteins to structural relationships through multiple convolutional layers. During the training process, supervised learning is performed using existing annotated data (such as proteins with known functions and the effects of their mutations on health) to optimize the parameters in the network. Finally, the trained genome function prediction model can effectively capture the relationship between protein structure and function, and output the genome function prediction results. For example, assuming that the input accepted by the model is the three-dimensional structure data of the genome protein, after being processed by the convolutional neural network, the output is the function prediction data of whether the protein has a chronic disease (such as diabetes or cancer). After the genome function prediction model is established, the next step is to input the target mapping genome coding region identification data into the model for target genome function prediction processing. The purpose of this step is to predict the potential physiological function changes caused by these mutations based on the characteristics of gene mutations and their relationship with protein function. In the specific implementation, the target mapping genome coding region identification data includes the site information of each gene mutation, the mutation type, and the risk category obtained by low-dimensional mapping. These data will be input into the pre-trained genome function prediction model. The model analyzes the input data and combines the three-dimensional structure and functional region information of the protein to predict the physiological function changes caused by the gene mutation. For example, if the target mapping genome coding region identification data shows that the mutation site of a gene (such as BRCA1) appears in the key area that inhibits tumor formation, the genome function prediction model predicts that the mutation is related to the occurrence of cancer. The target genome function prediction data output by the model usually includes the predicted function type, the risk score of dysfunction, and the potential impact of the variation on chronic diseases. Based on the target genome function prediction data, potential physiological indicator impact relationship analysis is performed to generate potential physiological indicator impact relationship data.The goal of this step is to analyze how gene variants affect potential physiological indicators (such as blood sugar levels, cholesterol concentrations, etc.) based on genome function prediction data and combined with physiological indicator data, so as to reveal the relationship between gene variants and chronic diseases. In the implementation process, the function prediction results and risk scores of each gene variant are obtained through the target genome function prediction data. Then, using known physiological indicators (such as biochemical examination data, physical examination data, etc.), combined with the functional prediction information of gene variants, a potential association model between gene variants and physiological indicators is established. This model will consider the impact of gene mutations on protein function, and predict how these gene variants affect physiological indicators based on the known relationship between the biological function of proteins and chronic diseases (such as diabetes, hypertension, etc.). For example, through analysis, some gene variants (such as TP53 mutations) are associated with the loss of cell repair function, which leads to the disorder of blood sugar regulation mechanism, thereby affecting the occurrence of diabetes. The output data of the model is the impact relationship data of potential physiological indicators, which contains the specific impact of gene variants on each physiological indicator and its potential impact on the occurrence of chronic diseases.

[0042] Furthermore, the individual chronic disease influencing factor analysis module includes the following functions: Performing historical pathology analysis on the electronic health record data to generate historical pathology analysis data, and performing historical pathology feature analysis based on the historical pathology analysis data to generate historical pathology feature data; Based on historical pathological characteristic data and individual physiological index data as individual pathological influencing factors, and using preset individual chronic disease composite interactive risk factor data to classify individual pathological influencing factors into pathological influencing factors of chronic disease types, individual chronic disease influencing factors are generated; The individual chronic disease impact factor evaluation algorithm is used to perform quantitative calculation and processing on the individual chronic disease impact factor to generate individual chronic disease impact factor quantitative data.

[0043] In an embodiment of the present invention, historical pathological analysis is performed on electronic health record (EHR) data. The specific operation is to extract all medical history information of the patient from the electronic health record, including but not limited to diagnosis information, treatment records, past medical history, drug use and related physiological data. For these data, the historical pathological information is analyzed by an algorithm to generate historical pathological analysis data, which reflects the types of diseases encountered by the patient in the past health process and their treatment effects. In order to ensure the accuracy of the analysis, a parsing algorithm based on natural language processing (NLP) is used to automatically extract information such as disease type, symptoms, treatment plan, etc. from the medical record text, and mark the corresponding time nodes and treatment results. After the historical pathological analysis data is generated, historical pathological feature analysis is performed, including pattern recognition and feature extraction of the parsed historical pathological data to reveal key pathological features related to chronic diseases. For example, for the historical pathological data of diabetic patients, analyze whether there are long-term hyperglycemia, islet dysfunction and other landmark features in their medical history. Through data mining technology, pathological features directly related to changes in the patient's health status are analyzed, and these features will provide a basis for subsequent chronic disease risk assessment. In a specific implementation, it is assumed that a patient's electronic health record contains a history of chronic hypertension. During the analysis process, the system extracts the patient's hypertension diagnosis information, medication records (such as the use of ACE inhibitors), and monitored blood pressure values ​​(such as higher than 140 / 90 mmHg all year round) for the past five years. These data are considered historical pathological analysis data. Further feature analysis found that there was a significant correlation between the patient's long-term hypertension and high-salt diet, and the characteristics of blood pressure fluctuations showed specific patterns, such as the association between the aggravation period and a specific season. The historical pathological feature data is combined with the individual physiological index data as individual pathological influencing factors. These data include but are not limited to the patient's medical history information, current physiological indicators (such as weight, blood sugar, blood lipids, etc.) and their historical changes. Combining historical pathological feature data and physiological index data, a series of pathological influencing factors are defined. For example, if a patient has long-term hyperglycemia and a family history of diabetes, then hyperglycemia and genetic factors are considered as potential pathological influencing factors. Then, the pathological influencing factors are classified by the preset individual chronic disease composite interactive risk factor data. This classification is completed through the established multi-factor interaction model, which comprehensively considers the impact of multiple factors such as genes, environment, and lifestyle on the patient's health. For example, the pathological influencing factors of diabetes type include hyperglycemia, insulin resistance, obesity, family history, etc. Based on the interaction of these factors, the classification algorithm will determine the patient's disease type based on different combinations of data. For example, if a patient's physiological data shows that his blood sugar level is abnormally elevated, and he is obese and has a family history, these factors will determine that the patient has a high risk of diabetes based on the interactive risk factor data, and thus be classified as a pathological influencing factor of diabetes.The individual chronic disease influencing factor evaluation algorithm is used to quantify the individual chronic disease influencing factors. First, based on the classification of pathological influencing factors, the evaluation algorithm will assign a weight to each influencing factor, which reflects the contribution of the factor to the occurrence of chronic diseases. The setting of weights can be determined through a large amount of clinical data analysis and statistical models, usually estimated using methods such as regression analysis, machine learning, or expert systems. For each influencing factor, such as hyperglycemia and obesity, the evaluation algorithm quantifies its degree of influence through a series of statistical indicators (such as risk ratio, standardized regression coefficient, etc.), and each pathological influencing factor will be assigned a quantitative value. After completing the quantitative calculation, the algorithm will integrate these quantitative data into a set of individual chronic disease influencing factor quantitative data. This data can be used as the basis for subsequent risk assessment and decision support to help doctors determine the chronic disease risk level of patients.

[0044] Furthermore, the individual chronic disease impact factor evaluation algorithm is as follows: ; ; ; ; In the formula, Expressed as the quantitative value of the individual chronic disease impact factor, It is expressed as the number of comprehensive influencing parameters of physiological indicators, Expressed as Comprehensive impact weight information of physiological indicators, Expressed as The impact value of chronic diseases on physiological indicators, Expressed as A time series sample set of chronic disease impact values ​​of physiological indicators, Expressed as The impact of chronic diseases on physiological indicators over time The standardized function of Expressed as the number of historical pathological characteristic parameters, Expressed as The weight information of historical pathological characteristic parameters, Expressed as The chronic disease impact value of the historical pathological characteristic parameters, Expressed as Historical pathological characteristic parameters, Expressed as The pathology-specific influence function of the historical pathological characteristic parameters, Expressed as The pathology-specific influence function adjustment coefficient of the historical pathological characteristic parameters, Represented as the chronic disease risk weight information of an individual’s genome, Expressed as the genomic impact value of chronic diseases, Chronic disease impact adjustment factors expressed as genomic transformations, Gene signature Individual physiological indicators Genetic and physiological factors of chronic diseases affect the conversion function.

[0045] In the implementation algorithm of the present invention, the quantified value of the individual chronic disease influencing factor is It reflects the quantitative values ​​of the influencing factors of various chronic diseases, and evaluates their comprehensive impact on the patient's risk of various chronic diseases by calculating the weighted sum of physiological indicators, historical pathological characteristics, and genomic data. Through time series analysis and standardized functions, the changing trends of physiological indicators are quantified and combined with weight coefficients to reflect the long-term impact of these physiological indicators on chronic diseases. Each physiological indicator needs to be analyzed through separate time series data, and its respective weight It will be set based on the extent of its impact on chronic diseases. The impact of historical pathological characteristics in the chronic disease risk is calculated by standardized pathological characteristics and adjustment coefficients. This part takes into account the individual's previous pathological conditions, such as whether they have ever suffered from related diseases, and weights them according to their role in the current chronic disease risk. The genomic information is finally quantified into the genetic impact value of chronic diseases through the interaction between genes and physiological data. This part reflects the combined effect of genetic susceptibility and external physiological factors, and provides the contribution of the genome to chronic diseases. The individual chronic disease influencing factor evaluation algorithm quantifies and evaluates the individual's chronic disease risk by comprehensively considering physiological indicators, historical pathological characteristics, genomic information and other factors. It integrates a variety of factors that affect the risk of chronic diseases, including physiological indicators (such as blood pressure, heart rate, etc.), historical pathological characteristics (such as past medical history, family history) and genomic data (such as genetic susceptibility). This integration of multi-source data makes the evaluation more comprehensive, thereby improving the accuracy of chronic disease risk prediction. By assigning different weights to each parameter and combining the individual's physiological, pathological and genetic characteristics, the algorithm can perform chronic disease risk assessment based on the individual's unique situation. This personalized assessment can better reflect each person's actual health status. Chronic disease impact value of physiological indicators The calculation of the time series data is considered, which can dynamically reflect the trend of physiological indicators over time and improve the ability to predict the risk of long-term chronic diseases. , revealing the interaction between genetic and physiological factors, and providing a more accurate basis for the individualized prevention and control of chronic diseases. The setting and adjustment of parameter weights can adjust the evaluation model according to the actual impact of different data, thus ensuring the flexibility and adaptability of the algorithm.

[0046] Furthermore, the individual chronic disease composite interactive risk factor analysis module includes the following functions: Conduct correlation feature analysis on individual chronic disease influencing factors to generate correlation feature data on individual chronic disease assessment factors; Design a preliminary knowledge graph of individual chronic disease impact factors based on the quantitative data of individual chronic disease impact factors, and generate a preliminary individual chronic disease impact assessment graph; Based on the individual chronic disease assessment factor correlation characteristic data, the preliminary individual chronic disease impact assessment map is processed with correlation relationship identification to generate a correlated individual chronic disease impact assessment map; Analyze the related promoting factors of chronic diseases based on the related individual chronic disease impact assessment map to generate data on related promoting factors of chronic diseases; Using the data of chronic disease-related promoting influencing factors, the corresponding graph nodes in the individual chronic disease impact assessment graph are analyzed and processed for the nonlinear composite interactive characteristics of chronic disease influencing factors to generate composite interactive characteristic data of chronic disease influencing factors; According to the composite interactive characteristic data of chronic disease influencing factors, the quantitative data of individual chronic disease influencing factors are analyzed and processed into composite interactive risk factors of individual chronic disease influencing factors to generate individual chronic disease composite interactive risk factor data.

[0047] As an example of the present invention, refer to Figure 3 As shown, Figure 1 The functional flow diagram of the individual chronic disease composite interactive risk factor analysis module in this example includes: S501: Perform correlation characteristic analysis on individual chronic disease influencing factors to generate correlation characteristic data of individual chronic disease assessment factors; In an embodiment of the present invention, correlation characteristics analysis is performed on individual chronic disease influencing factors to identify the mutual relationship between individual chronic disease influencing factors. Through a variety of data mining techniques, it is identified which influencing factors have a strong statistical correlation with each other. For example, in the assessment of diabetes, it is found that blood sugar levels are strongly correlated with insulin resistance, and there is also a significant correlation between hypertension and obesity. In specific implementation, individual pathological influencing factors (such as hyperglycemia, hypertension, obesity, etc.) are combined with relevant physiological data (such as blood sugar value, blood pressure value, body weight, etc.), and the Pearson correlation coefficient is used to calculate the linear relationship between different factors. Next, these factors are grouped by category, such as performing association degree grouping, and the degree of association between each group of factors is further analyzed. In this way, individual chronic disease assessment factor association characteristic data is obtained, which reveals the interaction between various pathological factors and provides a basis for subsequent analysis. For example, the chronic disease influencing factors of an individual include hyperglycemia, obesity, hypertension and bad eating habits. By calculating the correlation coefficient, it is obtained that the correlation between hyperglycemia and obesity is 0.85, and the correlation between obesity and hypertension is 0.78. Based on this, the generated individual chronic disease assessment factor association characteristic data are as follows: a strong correlation between hyperglycemia and obesity, and a moderate correlation between hypertension and obesity.

[0048] S502: Design a preliminary knowledge graph of individual chronic disease impact factors based on the quantitative data of individual chronic disease impact factors, and generate a preliminary individual chronic disease impact assessment graph; In an embodiment of the present invention, a preliminary knowledge graph design is performed based on the quantitative data of individual chronic disease influencing factors (such as quantitative values ​​such as blood sugar value, blood pressure value, and body mass index). The specific operation is to use a graph theory model to take individual chronic disease influencing factors as nodes in the graph, and the edges between the nodes represent the relationship between different influencing factors. The weight of the edge is determined by the correlation data or statistical analysis results obtained in the association feature analysis. For example, in the evaluation graph of diabetes, the nodes may include "hyperglycemia", "insulin resistance", "obesity", etc., and the weight of the edge may be set according to the correlation analysis results. If the correlation coefficient between "hyperglycemia" and "insulin resistance" is 0.85, the weight of the edge between the two is 0.85. In this way, an individual chronic disease impact assessment graph is preliminarily constructed, and the relationship between chronic disease influencing factors is visualized in the form of a graph. Taking diabetes as an example, the graph includes nodes such as "hyperglycemia", "insulin resistance", "obesity", and "genetic factors", and the nodes are connected by edges, and the weight of the edge represents the strength of the association between these factors. For example, the strength of the association between hyperglycemia and insulin resistance is 0.85, and the strength of the association between hyperglycemia and genetic factors is 0.6. The map shows the multifactorial influence of diabetes through these relationships.

[0049] S503: performing association relationship identification processing on the preliminary individual chronic disease impact assessment map based on the individual chronic disease assessment factor association characteristic data to generate an associated individual chronic disease impact assessment map; In an embodiment of the present invention, the individual chronic disease assessment factor association feature data is applied to the preliminary individual chronic disease impact assessment map to further identify the association relationship between each node. Through in-depth analysis of the association feature data, the edges in the map will be recalibrated to clarify which influencing factors are the main driving factors. This process uses statistical methods or machine learning algorithms (such as Bayesian networks, decision trees, etc.) to adjust the relationship in the map to ensure that the association relationship between each node reflects the strength of the causal effect. For example, in the diabetes assessment map, if it is found that "hyperglycemia" is not only strongly correlated with "insulin resistance", but also has a certain correlation with "genetic factors", then it will be adjusted according to the actual impact of these features, so that a more realistic causal relationship is reflected in the map, and the association features are converted into association frequency information through the Bayesian algorithm, and the association degree is calculated based on the association frequency information. For the influence of genetic factors, it is necessary to adjust the association degree and the direction of the edge in the map to clarify the potential driving effect of "genetic factors" on "hyperglycemia". In the diabetes map, based on the associated feature data, it was found that the correlation between "genetic factors" and "hyperglycemia" was weak, so in the associated individual chronic disease impact assessment map, the edge weight between "genetic factors" and "hyperglycemia" was 0.3, reflecting that the impact of genetic factors on hyperglycemia is relatively indirect. Through this processing, the generated associated individual chronic disease impact assessment map more accurately presents the correlation strength between various factors.

[0050] S504: Analyze the related promoting factors of chronic diseases according to the related individual chronic disease impact assessment map to generate data on related promoting factors of chronic diseases; In an embodiment of the present invention, based on the associated individual chronic disease impact assessment map, the associated promoting factors of chronic diseases are analyzed through the relationship between the nodes in the map. By calculating the interaction between the nodes in the map (i.e., chronic disease influencing factors), find out which factors play a driving role in the patient's chronic disease risk and mark them as promoting factors of chronic diseases, and extract the key factors that promote pathological changes by analyzing the weighted influence relationship between the nodes. For example, in the map of diabetes, if the association strength between "hyperglycemia" and "insulin resistance" is large, and the interaction between them accelerates the occurrence of diabetes, these two factors can be regarded as associated promoting factors of diabetes. Based on these factors, chronic disease associated promoting factor data is generated, including the specific value of each promoting factor (such as the degree of influence) and its role in pathological development. In the assessment of chronic diseases of diabetes, the analysis results show that the strong association between "obesity" and "hyperglycemia" indicates that obesity is a key promoting factor of diabetes. Through further analysis, it is concluded that the promoting effect of obesity on the onset of diabetes is 0.85.

[0051] S505: using the chronic disease-related promoting influencing factor data, the corresponding graph nodes in the individual chronic disease impact assessment graph are subjected to nonlinear composite interactive characteristic analysis of chronic disease influencing factors to generate composite interactive characteristic data of chronic disease influencing factors; In an embodiment of the present invention, the data of influencing factors associated with chronic diseases are used to identify the potential interactions between the nodes in the graph (i.e., influencing factors of chronic diseases). These influencing factors include, but are not limited to, body mass index (BMI), blood sugar level, blood pressure value, cholesterol, etc., and these data are analyzed by statistical models or machine learning algorithms. For example, for an individual, his BMI value is 30, his blood sugar value is 200 mg / dL, and his blood pressure is 160 / 100 mmHg. By combining the data of influencing factors associated with chronic diseases, it is found that these factors are not simply linear relationships, but have complex nonlinear relationships, and are chronic disease risk factors with strong interactions. Next, nonlinear modeling techniques, such as support vector machines (SVM), random forests (Random Forest) or deep neural networks (DNN), are used to model the interactive relationships of these influencing factors. These techniques can capture complex patterns that cannot be reflected by traditional linear models by identifying high-order interaction effects between factors. For example, the interactive effect between blood sugar and blood pressure is not just an additive relationship, but involves a threshold effect. When blood sugar and blood pressure rise at the same time, the risk of cardiovascular disease will rise sharply, which is difficult to capture by linear models. In the specific operation process, the data of individual chronic disease influencing factors are input into the trained nonlinear model to calculate and extract the interaction characteristics between each pair of factors. For example, the nonlinear interaction characteristics of blood sugar and BMI are a composite function, expressed as "blood sugar × BMI" plus "BMI^2" to consider the high-order interaction between BMI and blood sugar. In this way, the risk of chronic diseases of individuals can be comprehensively evaluated from multiple dimensions. After the analysis is completed, the composite interaction characteristic data of chronic disease influencing factors are generated. This data contains the interaction intensity between various chronic disease influencing factors and their composite effects under nonlinear modeling. For example, through nonlinear modeling, it is found that the interaction between high blood sugar and obesity has a particularly significant effect on the risk of heart disease, and the weight value of its interaction characteristics reaches 0.9, while the weights of each factor considered separately are 0.5 and 0.6 respectively. This feature data provides deep interaction information for the nodes in the atlas, which helps to further accurately evaluate the risk of chronic diseases of individuals. Finally, combined with these composite interaction feature data, the node relationship in the atlas is updated to reflect the more sophisticated and complex interactions between chronic disease influencing factors.

[0052] S506: Perform composite interactive risk factor analysis of individual chronic disease influencing factors on the individual chronic disease influencing factor quantified data according to the composite interactive characteristic data of chronic disease influencing factors to generate individual chronic disease composite interactive risk factor data.

[0053] In an embodiment of the present invention, quantitative data of chronic disease influencing factors of individuals are collected, and these data include physiological indicators and medical history information related to various chronic diseases, such as individual body mass index (BMI), blood sugar level, blood pressure value, cholesterol content, exercise habits, eating habits and family medical history. These factors will be converted into quantitative values ​​after data collection for subsequent risk assessment. These quantitative data are combined with the composite interactive feature data of chronic disease influencing factors previously generated by nonlinear composite interactive feature analysis. Composite interactive feature data usually contains the interaction effects between different chronic disease influencing factors. For example, the interaction between blood sugar level and body mass index is expressed as "blood sugar × BMI" and has a certain interaction weight. The interactive feature reflects the comprehensive impact of blood sugar and BMI on chronic disease risk when they act together. Using these composite interactive feature data, machine learning models (such as random forests, gradient boosting trees, deep neural networks, etc.) are used to perform risk factor analysis. The model calculates the composite interactive risk under each factor combination by inputting the quantitative data of chronic disease influencing factors and the interactive feature data. For example, when an individual's blood sugar level is 200 mg / dL and BMI is 30, the model calculates that the contribution of this interaction to the risk of cardiovascular disease is 0.75, and the contribution of the interaction between blood pressure and cholesterol to the risk is 0.65. When conducting risk factor analysis, the model not only considers the individual impact of each factor, but also comprehensively considers the interaction effects of multiple factors. For example, the interaction between high BMI and elevated blood sugar levels increases the risk of diabetes, while the interaction between BMI and high blood pressure increases the risk of cardiovascular disease. By modeling these interactions, the chronic disease risks of specific combinations can be identified and risk scores for each factor combination can be generated. Based on the analysis results, individual chronic disease composite interaction risk factor data are generated. These data reflect the overall risk level of an individual under the interaction of multiple chronic disease risk factors. The data are usually presented in the form of risk scores or weight values, reflecting the risk of an individual under different factor combinations. For example, if the interaction score between an individual's blood glucose level and BMI is 0.75, and the interaction score between blood pressure and cholesterol is 0.65, then the final individual chronic disease composite interactive risk factor data will integrate these factors and give an overall risk assessment value (e.g., 0.72), indicating the comprehensive risk faced by the individual under the interaction of all chronic disease factors.

[0054] Furthermore, the individual chronic disease risk assessment module includes the following functions: Monitor and process the data of composite interactive risk factors of individual chronic diseases to generate monitoring data of composite interactive risk factors of individual chronic diseases; Perform cluster analysis of chronic disease risk factors on monitoring data of individual chronic disease composite interactive risk factors to generate cluster data of individual chronic disease composite interactive risk factors; Perform interactive risk factor cluster outlier trend slope analysis based on individual chronic disease composite interactive risk factor cluster data to generate interactive risk factor cluster outlier trend slope data; Based on the preset chronic disease risk trend level threshold, the interactive risk factor cluster outlier trend slope data is processed by trend slope classification to generate interactive risk factor cluster outlier trend slope classification data; Individual chronic disease risk assessment is performed based on individual chronic disease composite interactive risk factor data and interactive risk factor cluster outlier trend slope grading data to generate individual chronic disease risk assessment data.

[0055] In an embodiment of the present invention, the individual chronic disease composite interactive risk factor data generated in the above steps are continuously monitored, and the monitoring results are saved in a database to obtain individual chronic disease composite interactive risk factor monitoring data. After completing the data monitoring, cluster analysis of chronic disease risk factors is performed next. Chronic disease risk factors are classified and processed by clustering algorithms (such as K-means, hierarchical clustering, etc.). The goal is to cluster and group individual monitoring data. Taking blood sugar and BMI as examples, blood sugar and BMI related data are clustered and grouped, and outliers of each cluster group are monitored. And according to the changing trend of these outliers, the trend slope analysis of the clustering outliers of the interactive risk factor is performed to evaluate whether the risk factor is changing, and whether the trend of the change is upward or downward. For example, in the individual chronic disease composite interactive risk factor data within 6 months of continuous monitoring, if the blood sugar level and BMI of the patient individual in a certain cluster continue to increase, and this change trend is significant, then the individual's blood sugar and BMI changes will be considered as outliers, and the trend slope of the patient will be calculated. The trend slope can be calculated by linear regression or other correlation analysis methods. If the trend slope of the individual is positive and greater than 0.1 in the overall monitoring interval or each monitoring node, it means that the risk of chronic disease is increasing. The trend slope of the outlier is graded by comparing with the preset chronic disease risk trend level threshold. These thresholds are set according to the clinical standards of chronic disease risk, with the purpose of dividing risk changes into different levels. Assume that these thresholds are: high risk (trend slope>0.1): indicates that the relevant chronic disease risk factors show a rapid upward trend and urgent intervention is needed; medium risk (0.05 <trend slope≤0.1): indicates that the chronic disease risk factors have a certain upward trend, but are not serious; low risk (trend slope≤0.05): indicates that the risk factors change slowly or tend to be stable, and the condition is at a lower risk level. For example, for a patient, if the trend slope of his blood sugar is 0.12, which exceeds the threshold of 0.1, the change trend of blood sugar is rated as "high risk"; if the trend slope of his blood pressure is 0.07, which falls within the range of 0.05 to 0.1, the change trend of blood pressure is rated as "medium risk". Through this grading process, the system classifies each monitored interactive risk factor into different risk levels according to its change rate, and then determines the patient's chronic disease risk status. Taking into account the individual's chronic disease composite interactive risk factor data and trend slope grading data, the patient's overall chronic disease risk is evaluated. According to the trend slope grading data of the outlier, the change trend of each risk factor is combined with its corresponding chronic disease influencing factor, and a comprehensive evaluation is performed based on the change trend of these indicators and the patient's historical pathological data, and finally the individual chronic disease risk assessment data is obtained.

[0056] Furthermore, the performing interactive risk factor clustering outlier trend slope analysis based on individual chronic disease composite interactive risk factor clustering data includes: Performing interactive risk factor cluster outlier analysis based on individual chronic disease composite interactive risk factor cluster data to generate interactive risk factor cluster outlier data; Perform cluster outlier trend analysis on the interactive risk factor cluster outlier data to generate interactive risk factor cluster outlier trend data; The cluster outlier trend slope analysis is performed on the interactive risk factor cluster outlier trend data to generate the interactive risk factor cluster outlier trend slope data.

[0057] In an embodiment of the present invention, outlier detection is performed on the data points in each cluster group. Outliers refer to data points that are significantly different from the cluster center or most of the data points. Outlier detection methods include statistical methods (such as Z-score) or distance-based methods (such as density accessibility clustering, DBSCAN). For example, if a patient's blood sugar changes are significantly higher than the entire cluster group, the system will mark the patient's blood sugar data as an outlier. Through this process, the system will generate interactive risk factor cluster outlier data, which indicates which data points belong to outliers in the clusters of various chronic disease risk factors. These outliers usually represent abnormal fluctuations in the patient's chronic disease risk factors, indicating potential health problems. Trend analysis is performed on the previously identified interactive risk factor cluster outlier data, and time series analysis is performed on the data of each outlier to observe the changing trend of the outlier. For example, if a patient's blood sugar has shown significant fluctuations in the past few months, the fluctuation will continue to increase or slow down in the next few months. In order to analyze the trend of these outliers, the system uses regression analysis (such as linear regression, curve fitting, etc.) and time series analysis methods (such as ARIMA model) to fit the changing trend of each outlier and identify the rate of change of each outlier (such as increase, decrease or remain unchanged). These analysis results will generate interactive risk factor cluster outlier trend data to reflect the temporal change trend of outliers. These trend data will help the system understand whether chronic disease risk factors are gradually worsening and provide support for subsequent risk assessment. The interactive risk factor cluster outlier trend data is further analyzed by calculating the trend slope of each outlier to quantify the rate of change of these trends. The trend slope reflects the rate of change of the patient's chronic disease risk factors in a specific time period, generating interactive risk factor cluster outlier trend slope data.

[0058] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0059] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A chronic disease risk assessment system, characterized in that: Includes the following modules: A multi-source physiological sign data collection module is used to collect multi-source physiological sign data of patients through a medical cloud service platform to generate multi-source physiological sign data, wherein the multi-source physiological sign data includes electronic health record data, medical physiological monitoring signals and genomics data; A physiological index analysis module is used to analyze physiological indexes according to medical physiological monitoring signals and generate physiological index data; A physiological indicator potential impact feature analysis module is used to perform potential physiological indicator impact correlation weight analysis on genomics data based on physiological indicator data, and generate potential physiological indicator impact correlation weight data; Using the potential physiological index impact association weight data to perform individual physiological index weighting processing on the physiological index data, so as to obtain individual physiological index data; The individual chronic disease influencing factor analysis module is used to analyze and process the pathological influencing factors of chronic disease types based on electronic health record data and individual physiological indicator data to generate individual chronic disease influencing factors; perform quantitative calculations on individual chronic disease influencing factors to generate quantitative data of individual chronic disease influencing factors; An individual chronic disease composite interactive risk factor analysis module is used to analyze and process the composite interactive risk factors of individual chronic disease influencing factors based on the quantitative data of individual chronic disease influencing factors, and generate individual chronic disease composite interactive risk factor data; The individual chronic disease risk assessment module is used to perform individual chronic disease risk assessment processing based on individual chronic disease composite interactive risk factor data to generate individual chronic disease risk assessment data.

2. The chronic disease risk assessment system according to claim 1, characterized in that: The physiological index analysis module includes the following functions: Performing time series characteristic analysis of medical physiological monitoring signals according to the medical physiological monitoring signals to generate time series characteristic data of the medical physiological monitoring signals; Physiological index conversion processing is performed according to the time series characteristic data of the medical physiological monitoring signal to generate physiological index data.

3. The chronic disease risk assessment system according to claim 1, characterized in that: The physiological index potential impact feature analysis module includes the following functions: Acquire genome reference data from a medical database, wherein the genome reference data includes genome sequence reference data and genome protein structure-function relationship data; Performing whole genome sequencing analysis on genomic data to extract coding region information of genomic data; using genome sequence reference data to annotate the coding region information of genomic data to generate gene coding region identification data; Performing genome coding region-specific low-dimensional mapping reconstruction processing on the gene coding region identification data to generate mapped genome coding region identification data; According to the physiological indicator data, the mapping genome coding region identification data is analyzed for the associated target mapping node of the physiological indicator to generate the target mapping genome coding region identification data; Based on the genome protein structure-function relationship data, the target mapping genome coding region identification data is analyzed for potential physiological index impact relationship, generating potential physiological index impact relationship data, and designing potential physiological index impact association weight data through the potential physiological index impact relationship data; The physiological indicator data are weighted by individual physiological indicator weights using the potential physiological indicator impact association weight data to obtain individual physiological indicator data.

4. The chronic disease risk assessment system according to claim 3, characterized in that: The specific low-dimensional mapping reconstruction process of the genome coding region on the gene coding region identification data comprises: Perform coding region specific analysis based on gene coding region identification data to generate coding region specific data; Perform coding region specific heterogeneous feature analysis based on coding region specific data to generate coding region specific heterogeneous feature data; Perform coding region-specific mapping relationship analysis based on coding region-specific heterogeneous feature data to generate coding region-specific mapping decisions; The genome coding region identification data is subjected to genome coding region-specific low-dimensional mapping reconstruction through coding region-specific mapping decision to generate mapped genome coding region identification data.

5. The chronic disease risk assessment system according to claim 3, characterized in that: The analysis of the potential physiological index impact relationship of the target mapped genome coding region identification data based on the genome protein structure-function relationship data includes: Based on the preset convolutional neural network algorithm and genome protein structure-function relationship data, the mapping relationship between the protein structure and function of the genome is analyzed to obtain a genome function prediction model; Transmitting the target mapped genome coding region identification data to the genome function prediction model for target genome function prediction processing to generate target genome function prediction data; The potential physiological index impact relationship analysis is performed based on the target genome function prediction data to generate potential physiological index impact relationship data.

6. The chronic disease risk assessment system according to claim 1, characterized in that: The individual chronic disease influencing factor analysis module includes the following functions: Performing historical pathology analysis on the electronic health record data to generate historical pathology analysis data, and performing historical pathology feature analysis based on the historical pathology analysis data to generate historical pathology feature data; Based on historical pathological characteristic data and individual physiological index data as individual pathological influencing factors, and using preset individual chronic disease composite interactive risk factor data to classify individual pathological influencing factors into pathological influencing factors of chronic disease types, individual chronic disease influencing factors are generated; The individual chronic disease impact factor evaluation algorithm is used to perform quantitative calculation and processing on the individual chronic disease impact factor to generate individual chronic disease impact factor quantitative data.

7. The chronic disease risk assessment system according to claim 6, characterized in that: The individual chronic disease impact factor evaluation algorithm is as follows: ； ； ； ； In the formula, Expressed as the quantitative value of the individual chronic disease impact factor, It is expressed as the number of comprehensive influencing parameters of physiological indicators, Expressed as Comprehensive impact weight information of physiological indicators, Expressed as The impact value of chronic diseases on physiological indicators, Expressed as A time series sample set of chronic disease impact values ​​of physiological indicators, Expressed as The impact of chronic diseases on physiological indicators over time The standardized function of Expressed as the number of historical pathological characteristic parameters, Expressed as The weight information of historical pathological characteristic parameters, Expressed as The chronic disease impact value of the historical pathological characteristic parameters, Expressed as Historical pathological characteristic parameters, Expressed as The pathology-specific influence function of the historical pathological characteristic parameters, Expressed as The pathology-specific influence function adjustment coefficient of the historical pathological characteristic parameters, Represented as the chronic disease risk weight information of an individual’s genome, Expressed as the genomic impact value of chronic diseases, Chronic disease impact adjustment factors expressed as genomic transformations, Gene signature Individual physiological indicators Genetic and physiological factors of chronic diseases affect the conversion function.

8. The chronic disease risk assessment system according to claim 1, characterized in that: The individual chronic disease composite interactive risk factor analysis module includes the following functions: Conduct correlation feature analysis on individual chronic disease influencing factors to generate correlation feature data on individual chronic disease assessment factors; Design a preliminary knowledge graph of individual chronic disease impact factors based on the quantitative data of individual chronic disease impact factors, and generate a preliminary individual chronic disease impact assessment graph; Based on the individual chronic disease assessment factor correlation characteristic data, the preliminary individual chronic disease impact assessment map is processed with correlation relationship identification to generate a correlated individual chronic disease impact assessment map; Analyze the related promoting factors of chronic diseases based on the related individual chronic disease impact assessment map to generate data on related promoting factors of chronic diseases; Using the data of chronic disease-related promoting influencing factors, the corresponding graph nodes in the individual chronic disease impact assessment graph are analyzed and processed for the nonlinear composite interactive characteristics of chronic disease influencing factors to generate composite interactive characteristic data of chronic disease influencing factors; According to the composite interactive characteristic data of chronic disease influencing factors, the quantitative data of individual chronic disease influencing factors are analyzed and processed into composite interactive risk factors of individual chronic disease influencing factors to generate individual chronic disease composite interactive risk factor data.

9. The chronic disease risk assessment system according to claim 1, characterized in that: The individual chronic disease risk assessment module includes the following functions: Monitor and process the data of composite interactive risk factors of individual chronic diseases to generate monitoring data of composite interactive risk factors of individual chronic diseases; Perform cluster analysis of chronic disease risk factors on monitoring data of individual chronic disease composite interactive risk factors to generate cluster data of individual chronic disease composite interactive risk factors; Perform interactive risk factor cluster outlier trend slope analysis based on individual chronic disease composite interactive risk factor cluster data to generate interactive risk factor cluster outlier trend slope data; Based on the preset chronic disease risk trend level threshold, the interactive risk factor cluster outlier trend slope data is processed by trend slope classification to generate interactive risk factor cluster outlier trend slope classification data; Individual chronic disease risk assessment is performed based on individual chronic disease composite interactive risk factor data and interactive risk factor cluster outlier trend slope grading data to generate individual chronic disease risk assessment data.

10. The chronic disease risk assessment system according to claim 9, characterized in that: The method of performing interactive risk factor clustering outlier trend slope analysis based on individual chronic disease composite interactive risk factor clustering data includes: Performing interactive risk factor cluster outlier analysis based on individual chronic disease composite interactive risk factor cluster data to generate interactive risk factor cluster outlier data; Perform cluster outlier trend analysis on the interactive risk factor cluster outlier data to generate interactive risk factor cluster outlier trend data; The cluster outlier trend slope analysis is performed on the interactive risk factor cluster outlier trend data to generate the interactive risk factor cluster outlier trend slope data.

Images