Arteriosclerosis risk prediction method and device based on machine learning

Abstract

The invention discloses an arteriosclerosis risk prediction method and device based on machine learning, and belongs to the technical field of machine learning, and the method comprises the steps: obtaining electronic medical record data of a target object, and carrying out the standardization processing; inputting the standardized electronic medical record data into a machine learning model based on a LightGBM algorithm to identify high-risk individuals of arteriosclerosis; the influence of each feature on the arteriosclerosis risk prediction result is quantified by combining the contribution values of the input features, and targeted health management suggestions are provided for patients. The arteriosclerosis risk prediction model constructed by learning a large amount of clinical data not only can improve the accuracy of arteriosclerosis risk prediction, but also can reveal the complex relationship between risk factors which are difficult to find by a traditional statistical method, so that the design and implementation of a personalized health management scheme are realized. Therefore, a more scientific basis is provided for preventing and treating arteriosclerosis.

Description

Technical Field

[0001] The present invention belongs to the field of machine learning technology, and more specifically, relates to a method and device for predicting arteriosclerosis risk based on machine learning. Background Art

[0002] Atherosclerosis is a disease characterized by thickening, hardening, and loss of elasticity of the arterial walls. It usually develops with age and is one of the main factors leading to cardiovascular disease. This pathological condition can affect medium and large arteries, including those in important areas such as the coronary arteries, cerebral arteries, and renal arteries, thereby increasing the risk of heart attack, stroke, and other serious health problems. Existing methods for assessing arteriosclerosis mainly include blood pressure measurement, blood lipid analysis, electrocardiogram (ECG), ultrasound imaging, computed tomography (CT) angiography, and pulse wave velocity (PWV) testing.

[0003] The development and progression of arteriosclerosis is the result of a combination of factors, including but not limited to high blood pressure, high cholesterol, smoking, obesity, diabetes, lack of exercise, and genetic factors. These risk factors promote lipid deposition in the arterial intima through complex physiological mechanisms, leading to changes in the arterial wall, including inflammation, fibrosis, and smooth muscle cell proliferation, ultimately causing stenosis or occlusion of the lumen.

[0004] However, due to the excessive number of factors that need to be considered, the existing technologies for predicting arteriosclerosis risk are often not accurate enough. Summary of the Invention

[0005] In response to the above-mentioned defects or improvement needs of the prior art, the present invention provides a method and device for predicting arteriosclerosis risk based on machine learning, which aims to solve the technical problem that arteriosclerosis risk prediction in the prior art is often not accurate enough.

[0006] To achieve the above objectives, according to one aspect of the present invention, a method for predicting arteriosclerosis risk based on machine learning is provided, comprising:

[0007] S1: Standardize the electronic medical record data of multiple target subjects;

[0008] S2: Screening out classification feature data related to arteriosclerosis from the normalized electronic medical record data and using them as samples in the training set; wherein the samples are all based on patients;

[0009] S3: using each sample in the training set as input and the diagnostic risk result of arteriosclerosis corresponding to each sample in the training set as output, training a machine learning model based on the LightGBM algorithm to obtain an arteriosclerosis risk prediction model;

[0010] S4: inputting the current characteristic data related to arteriosclerosis corresponding to the current patient into the arteriosclerosis risk prediction model to predict and obtain a corresponding arteriosclerosis risk assessment result;

[0011] S5: Calculating the contribution value of each parameter in the current feature data to the arteriosclerosis risk assessment result;

[0012] S6: Parameters whose contribution to the arteriosclerosis risk assessment result is greater than a preset value are regarded as key risk factors, and targeted health management suggestions are provided to the current patient based on the key risk factors.

[0013] Furthermore, S2 includes: screening out classification feature data related to arteriosclerosis from the standardized electronic medical record data, including: age, gender, body mass index, low-density lipoprotein cholesterol, estimated glomerular filtration rate, whether or not the patient has diabetes, whether or not the patient has smoking, whether or not the patient has hypertension, uric acid, triglycerides and high-sensitivity C-reactive protein.

[0014] Furthermore, the diagnostic risk results of arteriosclerosis corresponding to each sample are evaluated using pulse wave velocity; when PWV is less than 1400 cm / s, the degree of arteriosclerosis is classified as normal, and the classification value is represented by 0; when PWV is greater than or equal to 1400 cm / s, the degree of arteriosclerosis is classified as abnormal, and the classification value is represented by 1.

[0015] Furthermore, the S1 includes: removing duplicate records, correcting format inconsistencies, filling missing values, and processing abnormal values of the target object's electronic medical record data, and then uniformly encoding to obtain the standardized electronic medical record data.

[0016] Furthermore, the machine learning model based on the LightGBM algorithm is trained in S3, including:

[0017] During the training process, key model parameters of the machine learning model are adjusted, including the learning rate, the number of trees, the maximum depth of the tree, and the number of leaves, and the key model parameters are optimized through cross-validation techniques to prevent overfitting.

[0018] Furthermore, the S5 includes: calculating the SHAP value of each parameter in the current characteristic data to the arteriosclerosis risk assessment result as the contribution value, so as to quantify the influence of each parameter on the arteriosclerosis risk prediction result.

[0019] Furthermore, the health management suggestions in S6 include at least one of adjusting eating habits, increasing physical activities, and controlling blood pressure and blood sugar.

[0020] According to another aspect of the present invention, there is provided an apparatus for predicting arteriosclerosis risk based on machine learning, comprising:

[0021] A normalization module is used to normalize the electronic medical record data of multiple target subjects;

[0022] A screening module is used to screen out classification feature data related to arteriosclerosis from the normalized electronic medical record data and use them as samples in a training set; wherein the samples are all based on patients;

[0023] a training module for training a machine learning model based on the LightGBM algorithm using each sample in the training set as input and the diagnostic risk result of arteriosclerosis corresponding to each sample in the training set as output, thereby obtaining an arteriosclerosis risk prediction model;

[0024] A prediction module, configured to input current characteristic data related to arteriosclerosis corresponding to the current patient into the arteriosclerosis risk prediction model to predict and obtain a corresponding arteriosclerosis risk assessment result;

[0025] a calculation module, configured to calculate the contribution of each parameter in the current characteristic data to the arteriosclerosis risk assessment result;

[0026] The suggestion module is used to regard the parameter whose contribution value to the arteriosclerosis risk assessment result is greater than a preset value as a key risk factor, and provide targeted health management suggestions for the current patient based on the key risk factor.

[0027] According to another aspect of the present invention, an electronic device is provided, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above method when executing the computer program.

[0028] According to another aspect of the present invention, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above method are implemented.

[0029] In general, the above technical solutions conceived by the present invention can achieve the following beneficial effects compared with the prior art:

[0030] (1) The present invention provides a method for predicting the risk of arteriosclerosis based on machine learning, which obtains the electronic medical record data of the target object and performs normalization processing; inputs the normalized electronic medical record data into a machine learning model based on the LightGBM algorithm to identify individuals at high risk of arteriosclerosis; and quantifies the impact of each feature on the prediction result of arteriosclerosis risk by combining the contribution value of the input features, thereby providing patients with targeted health management suggestions. Among them, by learning a large amount of clinical data, such as vital signs, blood biochemical indicators, lifestyle information, etc., constructing a machine learning model based on the LightGBM algorithm can not only improve the accuracy of arteriosclerosis risk prediction, but also reveal the complex relationship between risk factors that are difficult to discover with traditional statistical methods, thereby realizing the design and implementation of personalized health management plans, thereby providing a more scientific basis for the prevention and treatment of arteriosclerosis. In addition, the arteriosclerosis risk prediction model constructed by the present invention can help identify people with potential arteriosclerosis risk but have not yet shown typical symptoms, and has good application prospects.

[0031] (2) By standardizing electronic medical record data and converting it into a structured preset format for storage, the consistency and comparability of the data are ensured, making cross-system data exchange more convenient and facilitating the integration of multi-source data to enhance the effectiveness of model training.

[0032] (3) By initially screening key features that are causally related to arteriosclerosis and combining them with the LightGBM algorithm for feature selection, model training, and optimization, we were able to predict the risk of arteriosclerosis in the target subjects. In particular, while maintaining a decision threshold of 0.5, we optimized model parameters and evaluation metrics (such as AUC) to focus on improving the recall rate of the model on the validation set, thereby more effectively identifying individuals at high risk of arteriosclerosis.

[0033] (4) The interpretability of the model was improved through SHAP analysis, the impact of each feature on the prediction results of arteriosclerosis risk was quantified, and visual operation software was produced to enable doctors and patients to clearly understand which lifestyle habits or physiological indicators significantly affect individual risk assessment, and thus provide patients with targeted health management recommendations. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 is a flow chart of the method for predicting arteriosclerosis risk based on machine learning provided in Example 1 of the present invention;

[0035] Figure 2 This is a schematic diagram of the structured JSON format of standardized electronic medical record data provided in Example 1 of the present invention;

[0036] Figure 3is a ROC curve diagram of the arteriosclerosis risk prediction model provided in Example 1 of the present invention;

[0037] Figure 4 This is a SHAP waterfall chart of the arteriosclerosis risk of a target subject provided by Example 1 of the present invention;

[0038] Figure 5 This is the risk assessment software interface of the arteriosclerosis risk prediction model provided in Example 1 of the present invention. DETAILED DESCRIPTION

[0039] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely for the purpose of explaining the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.

[0040] Example 1

[0041] This embodiment provides a method for predicting arteriosclerosis risk based on machine learning, such as Figure 1 As shown, it includes: S1: normalizing the electronic medical record data of multiple target objects; S2: screening the classification feature data related to arteriosclerosis from the normalized electronic medical record data and using it as samples in the training set; wherein the samples are all based on patients; S3: using each sample in the training set as input and the diagnostic risk result of arteriosclerosis corresponding to each sample in the training set as output, training a machine learning model based on the LightGBM algorithm to obtain an arteriosclerosis risk prediction model; S4: inputting the current feature data related to arteriosclerosis corresponding to the current patient into the arteriosclerosis risk prediction model to predict the corresponding arteriosclerosis risk assessment result; S5: calculating the contribution value of each parameter in the current feature data to the arteriosclerosis risk assessment result; S6: considering the parameter whose contribution value to the arteriosclerosis risk assessment result is greater than a preset value as a key risk factor, and providing targeted health management suggestions for the current patient based on the key risk factor. Wherein, the preset value is set based on the actual scenario, and actually multiple key risk factors are ranked from the largest to the smallest in contribution value.

[0042] The present invention provides a method for predicting the risk of arteriosclerosis based on machine learning, the method comprising: obtaining the electronic medical record data of the target subject and performing normalization processing; inputting the normalized electronic medical record data into an arteriosclerosis risk prediction model to identify individuals at high risk of arteriosclerosis; quantifying the impact of each feature on the arteriosclerosis risk prediction result in combination with the SHAP value of the input feature, and providing patients with targeted health management suggestions. Among them, by learning a large amount of clinical data, such as vital signs, blood biochemical indicators, lifestyle information, etc., constructing a machine learning model based on the LightGBM algorithm can help identify people who have potential arteriosclerosis risks but have not yet shown typical symptoms. The machine learning model based on the LightGBM algorithm can not only improve the accuracy of arteriosclerosis risk prediction, but also reveal the complex relationship between risk factors that are difficult to discover with traditional statistical methods, thereby realizing the design and implementation of personalized health management plans, thereby providing a more scientific basis for the prevention and treatment of arteriosclerosis.

[0043] As an optional implementation, S1 collects the target subject's electronic medical record data, including vital signs, blood biochemical indicators, and lifestyle information, and normalizes the electronic medical record data to ensure data consistency and comparability. As an optional implementation, S1 includes: removing duplicate records, correcting format inconsistencies, filling missing values, and processing outliers from the target subject's electronic medical record data, and then uniformly encoding the data to obtain the normalized electronic medical record data.

[0044] Specifically, in step S1, the normalization of electronic medical record data refers to a comprehensive preprocessing operation on the original data, including removing duplicate records, correcting format inconsistencies, filling missing values, and handling outliers. In addition, the features are uniformly encoded into a form suitable for model input, and the cleaned and processed data are finally converted into a structured JSON format for storage, so as to facilitate subsequent data analysis and modeling. Figure 2 shown.

[0045] As an optional embodiment, S2 includes: screening out classification feature data related to arteriosclerosis from the normalized electronic medical record data, including: age, gender, body mass index, low-density lipoprotein cholesterol, estimated glomerular filtration rate, diabetes, smoking status, hypertension, uric acid, triglycerides, and high-sensitivity C-reactive protein. As an optional embodiment, the diagnostic risk result of arteriosclerosis corresponding to each sample is assessed using pulse wave velocity; when PWV is less than 1400 cm / s, the degree of arteriosclerosis is classified as normal, and the classification value is represented by 0; when PWV is ≥1400 cm / s, the degree of arteriosclerosis is classified as abnormal, and the classification value is represented by 1.

[0046] Specifically, feature screening of standardized electronic medical record data refers to identifying and selecting features that are highly correlated with arteriosclerosis from the processed data. By training the LightGBM model and using the gain method to calculate the feature importance, the contribution of each feature in the model is measured, that is, the average gain brought by the feature in all splits. Then, based on the feature importance score, the most predictive feature subset is screened out. These screened features are then used for subsequent model training. This process can not only improve the performance and interpretability of the model, but also reduce unnecessary computational complexity and improve overall analysis efficiency. In addition, the feature importance ranking is displayed through visualization tools to help more intuitively understand which features are most critical for the prediction of arteriosclerosis.

[0047] For example, further feature screening was conducted on standardized electronic medical record data, which was divided into a training set and a test set in an 8:2 ratio. The training set data was used for model training, and the test set data was used to detect the model's generalization ability, accuracy, and robustness. By training the LightGBM model and using the gain method to calculate feature importance, the top ten feature variables in terms of feature importance were selected as key feature variables. Based on past electronic medical record data, the top ten feature variables in terms of feature importance were screened as follows: isHypertensive, eGFR, LDL-C, TG, BMI, UA, CRP, isDiabetic, Age, and isSmoker.

[0048] As an optional implementation, the S3 trains a machine learning model based on the LightGBM algorithm, including: adjusting key model parameters of the machine learning model during the training process, including the learning rate, the number of trees, the maximum depth of the tree, and the number of leaves, and optimizing the key model parameters through cross-validation technology to prevent overfitting.

[0049] Specifically, key parameters of the LightGBM model (such as the learning rate, number of trees, maximum tree depth, number of leaves, etc.) were adjusted and optimized through cross-validation techniques to prevent overfitting and improve the model's generalization ability. In particular, while maintaining a decision threshold of 0.5, by optimizing model parameters and evaluation metrics (such as AUC), the model's recall rate on the validation set was improved, thereby more effectively identifying individuals at high risk of arteriosclerosis and achieving risk prediction of arteriosclerosis in the target subjects.

[0050] For example, the features screened out in the training set are used as input variables, and the evaluation results of arteriosclerosis are used as output variables, and the LightGBM algorithm is used for model training. The key parameters of the LightGBM model (such as learning rate, number of trees, maximum depth of trees, number of leaves, etc.) are adjusted, and the optimal parameters are found through grid search (GridSearchCV). While keeping the decision threshold at 0.5, the model parameters and evaluation indicators (such as AUC) are optimized to focus on improving the recall rate (Recall) of the model on the validation set. These parameters are optimized through cross-validation technology to prevent overfitting and improve the generalization ability of the model. The optimal hyperparameter combination finally obtained is: the learning rate (learning_rate) is 0.01, the maximum depth (max_depth) is 10, the number of estimators (n_estimators) is 500, and the number of leaf nodes (num_leaves) is 31. Through the above content, the final arteriosclerosis risk prediction model can be obtained, which can more effectively identify high-risk individuals for arteriosclerosis and achieve risk prediction of arteriosclerosis for the target object. The ROC curve of the arteriosclerosis risk prediction model is as follows Figure 3 The accuracy of the risk prediction model was 0.789, the AUC was 0.858, and the recall rate of positive arteriosclerosis was 0.90.

[0051] As an optional implementation, the S5 includes: calculating the SHAP value of each parameter in the current feature data to the arteriosclerosis risk assessment result as the contribution value, so as to quantify the influence of each parameter on the arteriosclerosis risk prediction result.

[0052] Specifically, the interpretability of the arteriosclerosis risk prediction model is improved through SHAP (Shapley Additive Explanations) analysis. The specific approach is as follows: First, the SHAP value of each input feature of the trained LightGBM model is calculated to quantify the impact of each feature on the arteriosclerosis risk prediction results. Then, the importance of all features and their impact patterns on the model output are visualized using summary plot to identify the most critical risk factors. For specific patients, shap.plots.waterfall is used to display the specific impact of their personal data characteristics on the prediction of arteriosclerosis risk, so that doctors and patients can clearly understand which lifestyle habits or physiological indicators significantly affect individual risk assessments, and then provide patients with targeted health management suggestions, such as adjusting eating habits, increasing physical activity, controlling blood pressure and blood sugar, and other measures to reduce the risk of arteriosclerosis.

[0053] For example, the interpretability of the arteriosclerosis risk prediction model is improved through SHAP (Shapley Additive Explanations) analysis. The specific approach is to calculate and visualize the contribution of each feature to the prediction result (SHAP value), and identify the key risk factors whose contribution to the arteriosclerosis risk assessment result is greater than the preset value. Then, for the prediction results of a specific patient, the SHAP waterfall chart (shap.plots.waterfall) is used to explain in detail the impact of each feature, so as to provide patients with personalized health advice, such as lifestyle adjustments and preventive measures, to reduce the risk of arteriosclerosis. This step takes a certain target object as an example to carry out SHAP analysis of arteriosclerosis risk prediction. The corresponding feature variables are: eGFR (33.5) TG (3.24), BMI (24.7), LDL-C (3.43), UA (412.4), isHypertensive (1), isSmoker (0), CRP (20.5) Age (55), isDiabetic (1). Through SHAP analysis, the SHAP waterfall chart obtained is as follows: Figure 4 As shown in the figure, the two indicators that contribute more to the risk of arteriosclerosis for this target subject than the preset values are hypertension and glomerular filtration rate. Therefore, it can be recommended that this target subject pay attention to controlling blood pressure in life, maintaining blood pressure stability through a low-salt diet, regular exercise, and medication when necessary. At the same time, pay attention to kidney health, stay away from substances that are harmful to the kidneys, and regularly monitor kidney function, thereby effectively reducing the risk of arteriosclerosis.

[0054] As an optional implementation, the health management suggestions in S6 include at least one of adjusting eating habits, increasing physical activities, and controlling blood pressure and blood sugar.

[0055] Specifically, a visualization application is built based on the arteriosclerosis risk prediction model. The convenient operation software can help patients better use and understand the prediction process of arteriosclerosis risk, and enable patients to clearly understand which lifestyle habits or physiological indicators significantly affect individual risk assessment. The visualization interface provided in this embodiment is as follows: Figure 5 As shown, the left side is the electronic medical record data input window, the right side is provided with the interpretation of risk assessment and SHAP waterfall chart, and the bottom is provided with start assessment and exit buttons.

[0056] Example 2

[0057] This embodiment provides an arteriosclerosis risk prediction device based on machine learning, comprising: a normalization module, a screening module, a training module, a prediction module, a calculation module, and a suggestion module. The normalization module is configured to normalize the electronic medical record data of multiple target subjects; the screening module is configured to filter out classification feature data related to arteriosclerosis from the normalized electronic medical record data and use it as samples in a training set, wherein each sample is based on a patient; the training module is configured to train a machine learning model based on the LightGBM algorithm using each sample in the training set as input and the corresponding arteriosclerosis risk diagnosis result of each sample in the training set as output, thereby obtaining an arteriosclerosis risk prediction model; the prediction module is configured to input current feature data related to arteriosclerosis corresponding to the current patient into the arteriosclerosis risk prediction model to predict a corresponding arteriosclerosis risk assessment result; the calculation module is configured to calculate the contribution value of each parameter in the current feature data to the arteriosclerosis risk assessment result; and the suggestion module is configured to identify parameters whose contribution value to the arteriosclerosis risk assessment result is greater than a preset value as key risk factors, and provide targeted health management suggestions for the current patient based on the key risk factors.

[0058] Example 3

[0059] This embodiment provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above method when executing the computer program.

[0060] Example 4

[0061] This embodiment provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the above method are implemented.

[0062] It will be easily understood by those skilled in the art that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for predicting arteriosclerosis risk based on machine learning, characterized in that: include: S1: Standardize the electronic medical record data of multiple target subjects; S2: Screening out classification feature data related to arteriosclerosis from the normalized electronic medical record data and using them as samples in the training set; wherein the samples are all based on patients; S3: using each sample in the training set as input and the diagnostic risk result of arteriosclerosis corresponding to each sample in the training set as output, training a machine learning model based on the LightGBM algorithm to obtain an arteriosclerosis risk prediction model; S4: inputting the current characteristic data related to arteriosclerosis corresponding to the current patient into the arteriosclerosis risk prediction model to predict and obtain a corresponding arteriosclerosis risk assessment result; S5: Calculating the contribution value of each parameter in the current feature data to the arteriosclerosis risk assessment result; S6: Parameters whose contribution to the arteriosclerosis risk assessment result is greater than a preset value are regarded as key risk factors, and targeted health management suggestions are provided to the current patient based on the key risk factors.

2. The method for predicting arteriosclerosis risk based on machine learning according to claim 1, wherein: The S2 includes: screening out classification feature data related to arteriosclerosis from the standardized electronic medical record data, including: age, gender, body mass index, low-density lipoprotein cholesterol, estimated glomerular filtration rate, whether or not the patient is diabetic, whether or not the patient is smoking, whether or not the patient is hypertensive, uric acid, triglycerides and high-sensitivity C-reactive protein.

3. The method for predicting arteriosclerosis risk based on machine learning according to claim 2, wherein: The diagnostic risk results of arteriosclerosis corresponding to each sample are evaluated using pulse wave velocity; when PWV is less than 1400 cm / s, the degree of arteriosclerosis is classified as normal, and the classification value is represented by 0; when PWV is greater than or equal to 1400 cm / s, the degree of arteriosclerosis is classified as abnormal, and the classification value is represented by 1.

4. The method for predicting arteriosclerosis risk based on machine learning according to claim 1, wherein: Said S1 comprises: removing duplicate records, correcting format inconsistencies, filling missing values and processing abnormal values of the electronic medical record data of the target object, and then performing unified coding to obtain the standardized electronic medical record data.

5. The method for predicting arteriosclerosis risk based on machine learning according to claim 1, wherein: The S3 trains the machine learning model based on the LightGBM algorithm, including: adjusting the key model parameters of the machine learning model during the training process, including the learning rate, the number of trees, the maximum depth of the tree, and the number of leaves, and optimizing the key model parameters through cross-validation technology to prevent overfitting.

6. The method for predicting arteriosclerosis risk based on machine learning according to claim 5, wherein: The S5 includes: calculating the SHAP value of each parameter in the current feature data to the arteriosclerosis risk assessment result as the contribution value, so as to quantify the influence of each parameter on the arteriosclerosis risk prediction result.

7. The method for predicting arteriosclerosis risk based on machine learning according to claim 6, wherein: The health management suggestions in S6 include at least one of adjusting eating habits, increasing physical activities, and controlling blood pressure and blood sugar.

8. A device for predicting arteriosclerosis risk based on machine learning, characterized in that: include: A normalization module is used to normalize the electronic medical record data of multiple target subjects; A screening module is used to screen out classification feature data related to arteriosclerosis from the normalized electronic medical record data and use them as samples in a training set; wherein the samples are all based on patients; a training module for training a machine learning model based on the LightGBM algorithm using each sample in the training set as input and the diagnostic risk result of arteriosclerosis corresponding to each sample in the training set as output, thereby obtaining an arteriosclerosis risk prediction model; A prediction module, configured to input current characteristic data related to arteriosclerosis corresponding to the current patient into the arteriosclerosis risk prediction model to predict and obtain a corresponding arteriosclerosis risk assessment result; a calculation module, configured to calculate the contribution of each parameter in the current characteristic data to the arteriosclerosis risk assessment result; The suggestion module is used to regard the parameter whose contribution value to the arteriosclerosis risk assessment result is greater than a preset value as a key risk factor, and provide targeted health management suggestions for the current patient based on the key risk factor.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images