A method and system for constructing a cardiovascular disease risk prediction model

Abstract

The present invention relates to the technical field of model construction, and provides a method and system for constructing a cardiovascular disease risk prediction model, the method comprising: obtaining prediction parameters and prediction labels of a user; constructing a first risk prediction model for predicting the cardiovascular disease risk of the user; expanding the first risk prediction model to obtain a second risk prediction model; expanding the second risk prediction model to obtain a third risk prediction model; using the prediction parameters and prediction labels, training the first risk prediction model, the second risk model and the third risk prediction model to obtain a trained first risk prediction model, a trained second risk model and a trained third risk prediction model; constructing a target risk prediction model for predicting the risk of cardiovascular disease according to the trained first risk prediction model, the trained second risk model and the trained third risk prediction model. The present invention can improve the training efficiency of the cardiovascular disease risk prediction model.

Description

Technical Field

[0001] The present invention relates to the technical field of model construction, and in particular to a method, system and electronic equipment for constructing a cardiovascular disease risk prediction model. Background Art

[0002] Nowadays, cardiovascular disease is one of the leading causes of death and disease burden worldwide. Tens of millions of people die from cardiovascular disease every year, and 80% of these deaths can be effectively prevented through early identification. Currently, there are some cardiovascular disease prediction tools, such as the Framingham Cardiovascular Disease Risk Prediction Website and the Cardiovascular Disease Risk Assessment Website, which use multiple biomarkers to predict the risk of cardiovascular disease.

[0003] However, with the increasing social pressure, the number of young people suffering from cardiovascular disease has increased year by year, but the existing prediction tools are not suitable for the contemporary young population. First, the predictive factors used are numerous and complex, and young people usually do not have much time to go to the hospital for testing, which increases the difficulty of communicating with the subjects and model training. Similarly, some individuals may not be able to obtain comprehensive biomarker testing.

[0004] Therefore, it is particularly important to construct a cardiovascular disease risk prediction model suitable for young people. Such a model should be simpler, more accurate, easier to train, and able to be effectively applied in clinical and public health practices for young patients and potential patients. Summary of the invention

[0005] In view of the technical problems existing in the prior art, the present invention provides a method, system and electronic equipment for constructing a cardiovascular disease risk prediction model, which can improve the training efficiency of the cardiovascular disease risk prediction model.

[0006] The technical solution of the present invention to solve the above technical problems is as follows:

[0007] The present invention provides a method for constructing a cardiovascular disease risk prediction model, the method comprising:

[0008] Get the user's prediction parameters and prediction labels;

[0009] Constructing a first risk prediction model for predicting the cardiovascular disease risk of the user;

[0010] Expanding the first risk prediction model to obtain a second risk prediction model;

[0011] Expanding the second risk prediction model to obtain a third risk prediction model;

[0012] Using the prediction parameters and the prediction labels, training the first risk prediction model, the second risk model, and the third risk prediction model to obtain a trained first risk prediction model, a trained second risk model, and a trained third risk prediction model;

[0013] A target risk prediction model for predicting the risk of cardiovascular disease is constructed based on the trained first risk prediction model, the trained second risk model and the trained third risk prediction model.

[0014] Furthermore, the step of constructing a first risk prediction model for predicting the cardiovascular disease risk of the user includes:

[0015] Obtaining a first weight of each of the prediction parameters;

[0016] Define the weighted summation of the characteristics of each of the prediction parameters using the first weight to obtain the model construction requirements of the corresponding first risk prediction score, construct the model parameters of the first risk prediction model, and obtain the first risk prediction model.

[0017] Furthermore, the first risk prediction model is expanded to obtain a second risk prediction model, including:

[0018] Obtaining a second weight of each prediction parameter;

[0019] Acquire a first extension item for extending the first risk prediction model according to the second weight and the prediction parameter of the existing type;

[0020] Define the weighted summation of the characteristics of each of the prediction parameters and the first extension item using the second weight to obtain the model construction requirements of the corresponding second risk prediction score, construct the model parameters of the second risk prediction model, and obtain the second risk prediction model.

[0021] Furthermore, the second risk prediction model is expanded to obtain a third risk prediction model, including:

[0022] Obtaining a third weight of each prediction parameter;

[0023] Obtaining a second extension item for extending the second risk prediction model according to the third weight and the prediction parameter of the existing type;

[0024] Define the weighted summation of the characteristics of each of the prediction parameters and the second extension item using the third weight to obtain the model construction requirements of the corresponding third risk prediction score, construct the model parameters of the third risk prediction model, and obtain the third risk prediction model.

[0025] Furthermore, constructing a target risk prediction model for predicting cardiovascular disease risk based on the trained first risk prediction model, the trained second risk model and the trained third risk prediction model includes:

[0026] Get the fourth weight;

[0027] Define the fourth weight to perform weighted summation on the output results of the trained first risk prediction model, the trained second risk model and the trained third risk prediction model to obtain the model construction requirements of the corresponding target risk prediction score, construct the model parameters of the target risk prediction model, and obtain the target risk prediction model.

[0028] The present invention also provides a system for constructing a cardiovascular disease risk prediction model, the system comprising:

[0029] Data acquisition module, used to obtain the user's prediction parameters and prediction labels;

[0030] A first building module, used to build a first risk prediction model for predicting the cardiovascular disease risk of the user;

[0031] A second building module is used to expand the first risk prediction model to obtain a second risk prediction model;

[0032] A third building module is used to expand the second risk prediction model to obtain a third risk prediction model;

[0033] A model training module, used to train the first risk prediction model, the second risk model and the third risk prediction model using the prediction parameters and the prediction labels to obtain a trained first risk prediction model, a trained second risk model and a trained third risk prediction model;

[0034] The fourth construction module is used to construct a target risk prediction model for predicting the risk of cardiovascular disease based on the trained first risk prediction model, the trained second risk model and the trained third risk prediction model.

[0035] In addition, to achieve the above-mentioned purpose, the present invention also proposes an electronic device, comprising: a memory for storing a computer software program; a processor for reading and executing the computer software program, thereby realizing a method for constructing a cardiovascular disease risk prediction model as described above.

[0036] The beneficial effects of the present invention are:

[0037] (1) The target risk prediction model of the present invention uses only a relatively small number of biomarkers (such as the user's age, gender, blood pressure, blood sugar, cholesterol level, smoking status, etc.), which simplifies the input requirements and training requirements of the risk prediction model, and is convenient for promotion and application in clinical and public health practice. At the same time, the simplicity of the model reduces the difficulty of communication with the subjects, making it more suitable for young people with less free time, and improving the understanding and acceptance of the model results by clinicians and patients.

[0038] (2) In the process of constructing the target risk prediction model, the present invention can comprehensively consider multiple key biomarkers and the relationship between them (such as the ratio of total cholesterol to high-density lipoprotein cholesterol) without introducing too many prediction parameters. The constructed model can provide more accurate cardiovascular disease risk prediction. Without increasing the types of prediction parameters, complex nonlinear terms and interaction terms (such as the interaction between age and gender, the relationship between systolic blood pressure and diastolic blood pressure, etc.) are introduced to further improve the prediction performance and accuracy of the model.

[0039] (3) During the model training process, the present invention can flexibly adjust the model parameters of each sub-model according to the specific application scenario to meet the needs of different regions and populations. Each sub-model can be continuously optimized and updated to improve its long-term effectiveness and adaptability.

[0040] In summary, the present invention constructs a simple, accurate and fair risk prediction model for cardiovascular disease suitable for young people, which not only improves the prediction performance and training efficiency of the model, but also has important value in clinical application and public health practice. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 A scenario diagram of a method for constructing a cardiovascular disease risk prediction model provided by the present invention;

[0042] Figure 2 A flowchart of a method for constructing a cardiovascular disease risk prediction model provided by the present invention;

[0043] Figure 3 A schematic diagram of the structure of a system for constructing a cardiovascular disease risk prediction model provided by the present invention;

[0044] Figure 4 A schematic diagram of the hardware structure of a possible electronic device provided by the present invention. DETAILED DESCRIPTION

[0045] See also Figure 1 , Figure 1 A scenario diagram of a method for constructing a cardiovascular disease risk prediction model provided by the present invention. Figure 1As shown, the terminal and the server are connected via a network, such as a wired or wireless network connection. The terminal may include but is not limited to portable terminals such as mobile phones and tablets installed with various network platform applications, as well as fixed terminals such as computers, query machines, and advertising machines. The server provides users with various business services, including service push servers, user recommendation servers, etc.

[0046] It should be noted that Figure 1 The scenario diagram of a method for constructing a cardiovascular disease risk prediction model shown is only an example. The terminal, server, and application scenario described in the embodiment of the present invention are intended to more clearly illustrate the technical solution of the embodiment of the present invention, and do not generate limitations on the technical solution provided by the embodiment of the present invention. Ordinary technicians in this field can know that with the evolution of the system and the emergence of new business scenarios, the technical solution provided by the embodiment of the present invention is also applicable to similar technical problems.

[0047] Among them, the terminal can be used for:

[0048] Get the user's prediction parameters and prediction labels;

[0049] Constructing a first risk prediction model for predicting the cardiovascular disease risk of the user;

[0050] Expanding the first risk prediction model to obtain a second risk prediction model;

[0051] Expanding the second risk prediction model to obtain a third risk prediction model;

[0052] Using the prediction parameters and the prediction labels, training the first risk prediction model, the second risk model, and the third risk prediction model to obtain a trained first risk prediction model, a trained second risk model, and a trained third risk prediction model;

[0053] A target risk prediction model for predicting the risk of cardiovascular disease is constructed based on the trained first risk prediction model, the trained second risk model and the trained third risk prediction model.

[0054] See also Figure 2 , provides a flowchart of a method for constructing a cardiovascular disease risk prediction model of the present invention, comprising the following steps:

[0055] Step 201: Obtain the user's prediction parameters and prediction labels.

[0056] Among them, the prediction parameters are various characteristic values ​​of the user used for model input and prediction. In some embodiments, the prediction parameters may include the age of the user to be predicted (A): the actual age of the user. Gender (G): the gender of the user, usually 0 represents female and 1 represents male. Systolic blood pressure (SBP): the systolic blood pressure value of the user, usually obtained by a blood pressure measuring instrument. Diastolic blood pressure (DBP): the diastolic blood pressure value of the user, usually obtained by a blood pressure measuring instrument. Blood sugar (BG): the blood sugar level of the user, usually obtained by blood testing. Total cholesterol (TC): the total cholesterol level of the user, usually obtained by blood testing. High-density lipoprotein cholesterol (HDL-C): the high-density lipoprotein cholesterol level of the user, usually obtained by blood testing. Smoking status (S): the smoking status of the user, which can be obtained through a questionnaire or user self-report, with 0 representing non-smoking and 1 representing smoking.

[0057] In some embodiments, doctors or medical institutions obtain the user's physiological and lifestyle data through routine physical examinations and interviews. Users can fill in their age, gender, smoking status, etc. through health management applications or questionnaires.

[0058] In some embodiments, the prediction label can be whether the user of the prediction parameter develops cardiovascular disease, for example, the label can be a vector 0 or 1. In some embodiments, the collected user feature data can be matched with the follow-up records to ensure that the data of each user is complete. The data is sorted and cleaned to remove incomplete or inaccurate records. Corresponding labels can be assigned based on whether the user has a cardiovascular event during the follow-up period. Label data is usually stored in a database or table corresponding to the user feature data to ensure data consistency and availability. Ultimately, a complete data set can be constructed for training and testing models. The data set includes the user's prediction parameters and prediction labels:

[0059] Feature matrix (X): Each row represents a user and each column represents a feature (age, gender, systolic blood pressure, diastolic blood pressure, blood glucose, total cholesterol, HDL cholesterol, smoking status, ratio of total cholesterol to HDL cholesterol).

[0060] Label vector (y): Each row represents a user’s label (0 or 1), indicating whether cardiovascular disease occurs.

[0061] It should be noted that the characteristic value of each prediction parameter of the present invention can be the original value of the prediction parameter, such as the original value of blood glucose is 5, etc., or it can be a normalized value of the prediction parameter, such as 0.5, 0.6, etc., or it can be a specific setting, which is not limited in this embodiment.

[0062] Step 202: construct a first risk prediction model for predicting the cardiovascular disease risk of the user.

[0063] In some embodiments, step 202 may include:

[0064] Obtaining a first weight of each of the prediction parameters;

[0065] Define the weighted summation of the characteristics of each of the prediction parameters using the first weight to obtain the model construction requirements of the corresponding first risk prediction score, construct the model parameters of the first risk prediction model, and obtain the first risk prediction model.

[0066] In some embodiments, the first risk prediction score may be expressed as:

[0067]

[0068] Among them, R is the first risk prediction score, A is age, G is gender, SBP is systolic blood pressure, DBP is diastolic blood pressure, BG is blood glucose, TC is total cholesterol, HDL-C is high-density lipoprotein cholesterol, S is smoking status, and β is the first weight of each prediction parameter.

[0069] In the specific implementation, A is age, usually in years; G is gender, usually represented by 0 and 1 (such as 0 for female and 1 for male), because gender has a significant impact on the risk of cardiovascular disease; SBP is systolic blood pressure, which refers to the blood pressure in the arteries when the heart contracts, usually measured in millimeters of mercury (mmHg). Higher systolic blood pressure is an important risk factor for cardiovascular disease; DBP is diastolic blood pressure, which refers to the blood pressure in the arteries when the heart relaxes, usually measured in millimeters of mercury (mmHg). Higher diastolic blood pressure is also a risk factor for cardiovascular disease. BG is blood glucose, which refers to the glucose concentration in the blood, usually measured in milligrams per deciliter (mg / dL) or millimoles per liter (mmol / L). Blood glucose level is a key indicator for predicting cardiovascular disease; TC is total cholesterol, which refers to the total cholesterol level in the blood, usually measured in milligrams per deciliter (mg / dL) or millimoles per liter (mmol / L). A higher total cholesterol level is a risk factor for cardiovascular disease; HDL-C is high-density lipoprotein cholesterol, which refers to the high-density lipoprotein cholesterol level in the blood, usually measured in milligrams per deciliter (mg / dL) or millimoles per liter (mmol / L). Higher HDL-C levels are generally considered to be protective because it can help reduce atherosclerosis; S is smoking status, which refers to whether an individual smokes, usually represented by 0 and 1 (such as 0 for non-smoking and 1 for smoking). Smoking is a significant risk factor for cardiovascular disease; β is the first weight of each parameter, which needs to be determined through model training.

[0070] In the specific implementation, The ratio of total cholesterol to high-density lipoprotein cholesterol is often used as an indicator of cardiovascular health. This ratio reflects one of the risk factors for atherosclerosis and cardiovascular disease risk, and directly reflects the ratio of "bad" cholesterol to "good" cholesterol in the body. A higher ratio indicates a higher total cholesterol relative to high-density lipoprotein cholesterol, which is generally considered an indicator of increased risk of cardiovascular disease. Generally speaking, if this ratio is high, then the risk of cardiovascular disease is also increased accordingly.

[0071] In the first risk prediction model, Adding it to the model as a predictor can help improve the accuracy of the model for individual cardiovascular disease risk. By adjusting the parameter β7, the model can predict the risk of cardiovascular disease based on the degree of change in this ratio.

[0072] Step 203: Expand the first risk prediction model to obtain a second risk prediction model.

[0073] In some embodiments, step 203 may include:

[0074] Obtaining a second weight of each prediction parameter;

[0075] Acquire a first extension item for extending the first risk prediction model according to the second weight and the prediction parameter of the existing type;

[0076] Define the weighted summation of the characteristics of each of the prediction parameters and the first extension item using the second weight to obtain the model construction requirements of the corresponding second risk prediction score, construct the model parameters of the second risk prediction model, and obtain the second risk prediction model.

[0077] In some embodiments, the second risk prediction score is expressed as:

[0078]

[0079] Among them, R' is the second risk prediction score, A is age, G is gender, SBP is systolic blood pressure, DBP is diastolic blood pressure, BG is blood glucose, TC is total cholesterol, HDL-C is high-density lipoprotein cholesterol, S is smoking status, γ is the second weight of each prediction parameter, γ9(A×G), γ 11 log(BG+1), γ 13 S 2 The first expansion item.

[0080] In the specific implementation, the term γ9(A×G) is the interaction between age and gender, which takes into account the possible impact of the combination of age and gender on the risk of cardiovascular disease. is the square root of the product of systolic and diastolic blood pressures, capturing the complex relationship between the two. 11 log(BG+1) is the logarithmic transformation of blood glucose, which can better handle the distribution characteristics of blood glucose values. is the inverse of the ratio of total cholesterol to high-density lipoprotein cholesterol, which takes into account the impact of this ratio on cardiovascular disease risk, γ 13 S 2 is the square of smoking status, which represents the possible nonlinear effect of smoking on cardiovascular disease risk.

[0081] In this way, the risk prediction model is expanded after calculating the first risk score. Without adding additional parameter types, the first extension item is added to the model parameters of the first risk prediction model. This can more comprehensively consider the complex relationship between various biomarkers and improve the accuracy and applicability of cardiovascular disease risk assessment. The construction of the model requires a large amount of training data and appropriate optimization methods to ensure that the model can effectively predict the risks of different individuals.

[0082] Step 204: Expand the second risk prediction model to obtain a third risk prediction model.

[0083] In some embodiments, step 204 may include:

[0084] Obtaining a third weight of each prediction parameter;

[0085] Obtaining a second extension item for extending the second risk prediction model according to the third weight and the prediction parameter of the existing type;

[0086] Define the weighted summation of the characteristics of each of the prediction parameters and the second extension item using the third weight to obtain the model construction requirements of the corresponding third risk prediction score, construct the model parameters of the third risk prediction model, and obtain the third risk prediction model.

[0087] In some embodiments, the third risk prediction score may be expressed as:

[0088]

[0089] Among them, R” is the third risk prediction score, A is age, G is gender, SBP is systolic blood pressure, DBP is diastolic blood pressure, BG is blood glucose, TC is total cholesterol, HDL-C is high-density lipoprotein cholesterol, S is smoking status, δ is the third weight of each prediction parameter, δ 10log(TC+1),δ 11 (G×S), δ 14 sin(A),δ 15 cos(G) is the second expansion term.

[0090] In the specific implementation, R” is the third risk score, A is age, G is gender, SBP is systolic blood pressure, DBP is diastolic blood pressure, BG is blood glucose, TC is total cholesterol, HDL-C is high-density lipoprotein cholesterol, S is smoking status, and δ is the third weight of each parameter.

[0091] In the specific implementation, δ 10 log(TC+1) is the logarithmic transformation of total cholesterol, which can better handle the distribution characteristics of total cholesterol. 11 (G×S) is the interaction between gender and smoking status, which represents the possible compound effect of the combination of gender and smoking status on the risk of cardiovascular disease. It is the square of the ratio of systolic pressure to diastolic pressure. This nonlinear term captures the complex relationship between blood pressure. It is the exponential conversion of blood sugar, which can correct the distribution characteristics of blood sugar value. 14 sin(A) and δ 15 cos(G) is the trigonometric transformation of age and gender, which takes into account the cyclical effects of age and gender on cardiovascular disease risk.

[0092] In this way, the risk prediction model is expanded after calculating the second risk score. Without adding additional parameter types, the second extension term is added to the model parameters of the second risk prediction model, and more nonlinear terms and interactions are introduced, in an attempt to more comprehensively capture the complex relationship between different variables to improve the accuracy of cardiovascular disease risk assessment. Each δ represents the contribution of the corresponding variable or combination of variables to the risk score R", and these coefficients need to be estimated through training data.

[0093] Step 205: Use the prediction parameters and the prediction labels to train the first risk prediction model, the second risk model, and the third risk prediction model to obtain a trained first risk prediction model, a trained second risk model, and a trained third risk prediction model.

[0094] In some embodiments, prediction parameters and prediction labels for training may be prepared in advance. Specifically, they include:

[0095] Prediction parameters: various physiological and behavioral data of the user (age, gender, systolic blood pressure, diastolic blood pressure, blood glucose, total cholesterol, high-density lipoprotein cholesterol, smoking status, and composite features such as the ratio of total cholesterol to high-density lipoprotein cholesterol).

[0096] Prediction label: a binary label indicating whether the user has a cardiovascular event during the follow-up period (0 means no event, 1 means event).

[0097] Furthermore, a training data set (feature matrix X and label vector y) can be constructed: the feature matrix X contains the prediction parameters of all users, and the label vector y contains the prediction labels of all users.

[0098] In some embodiments, the training process of the first risk prediction model, the second risk prediction model, and the third risk prediction model is as follows:

[0099] (1) Define the model: linear regression model;

[0100] (2) Data segmentation: Divide the dataset into training set and test set (for example, 80% as training set and 20% as test set);

[0101] (3) Model training: using the training set data, a linear regression algorithm is used to estimate the first weight, the second weight, and the third weight;

[0102] (4) Model evaluation: Evaluate model performance on the test set and calculate indicators such as accuracy, recall, and F1 score;

[0103] (5) Model optimization: Adjust model parameters or feature selection based on the evaluation results until the model performance reaches the expected level.

[0104] When deploying the model, the trained model can be deployed to the clinical application system or health management platform, new user prediction parameters can be input, the target risk prediction score can be calculated, and personalized health management recommendations and preventive measures can be provided to users based on the scoring results.

[0105] Through the above steps, multiple cardiovascular disease risk prediction models can be constructed, trained, and applied, and integrated into an efficient risk assessment tool to provide accurate cardiovascular disease risk prediction.

[0106] Step 206: construct a target risk prediction model for predicting cardiovascular disease risk based on the trained first risk prediction model, the trained second risk prediction model and the trained third risk prediction model.

[0107] In some embodiments, step 206 may include:

[0108] Get the fourth weight;

[0109] Define the fourth weight to perform weighted summation on the output results of the trained first risk prediction model, the trained second risk model and the trained third risk prediction model to obtain the model construction requirements of the corresponding target risk prediction score, construct the model parameters of the target risk prediction model, and obtain the target risk prediction model.

[0110] In some embodiments, the comprehensive risk prediction score is expressed as:

[0111] R 综合 =α1R+α2R′+α3R″;

[0112] Among them, R 综合 is the comprehensive risk score, R is the first risk prediction score, R' is the second risk prediction score, R" is the third risk prediction score, and α is the fourth weight.

[0113] In the specific implementation, R 综合 is the comprehensive risk score, R is the first risk score, R' is the second risk score, R" is the third risk score, and α is the fourth weight.

[0114] In the specific implementation, R 综合 The information of three different models is combined, and the different contributions of each model to the assessment of cardiovascular disease risk are comprehensively considered. By setting an appropriate fourth weight α, the individual's cardiovascular disease risk can be predicted more accurately, and the advantages and information of each model can be utilized as much as possible. Different weight settings can adapt to the circumstances of different individuals, such as paying attention to the influence of age factors or the contribution of nonlinear relationships. In practical applications, the selection of α1, α2, and α3 may need to be determined through cross-validation or other evaluation methods to ensure the generalization ability and prediction accuracy of the model on different data sets.

[0115] In some embodiments, the trained target risk prediction model can be used to make predictions based on the prediction parameters of different users to obtain the target risk prediction score R corresponding to the user. 综合 , it is understandable that if the score R 综合 If the value of is greater than a certain threshold, it means that the user has a high probability of suffering from cardiovascular disease and needs to be prevented and checked. In addition, the target risk prediction score predicted by the target risk prediction model of the present invention is only used for reference and is not directly used as a basis for diagnosis and treatment of the user.

[0116] See also Figure 3 , Figure 3 A schematic diagram of the structure of a system for constructing a cardiovascular disease risk prediction model provided by the present invention.

[0117] like Figure 3As shown, a system for constructing a cardiovascular disease risk prediction model proposed in an embodiment of the present invention includes:

[0118] Data acquisition module 301, used to obtain prediction parameters and prediction labels of users;

[0119] A first building module 302 is used to build a first risk prediction model for predicting the cardiovascular disease risk of the user;

[0120] A second construction module 303 is used to expand the first risk prediction model to obtain a second risk prediction model;

[0121] A third construction module 304 is used to expand the second risk prediction model to obtain a third risk prediction model;

[0122] A model training module 305 is used to train the first risk prediction model, the second risk model and the third risk prediction model using the prediction parameters and the prediction labels to obtain a trained first risk prediction model, a trained second risk model and a trained third risk prediction model;

[0123] The fourth construction module 306 is used to construct a target risk prediction model for predicting the risk of cardiovascular disease based on the trained first risk prediction model, the trained second risk model and the trained third risk prediction model.

[0124] See also Figure 4 , Figure 4 Schematic diagram of an electronic device provided by an embodiment of the present invention. Figure 4 As shown, an embodiment of the present invention provides an electronic device 400, including a memory 410, a processor 420, and a computer program 411 stored in the memory 410 and executable on the processor 420. When the processor 420 executes the computer program 411, the following steps are implemented:

[0125] Get the user's prediction parameters and prediction labels;

[0126] Constructing a first risk prediction model for predicting the cardiovascular disease risk of the user;

[0127] Expanding the first risk prediction model to obtain a second risk prediction model;

[0128] Expanding the second risk prediction model to obtain a third risk prediction model;

[0129] Using the prediction parameters and the prediction labels, training the first risk prediction model, the second risk model, and the third risk prediction model to obtain a trained first risk prediction model, a trained second risk model, and a trained third risk prediction model;

[0130] A target risk prediction model for predicting the risk of cardiovascular disease is constructed based on the trained first risk prediction model, the trained second risk model and the trained third risk prediction model.

[0131] It should be noted that in the above embodiments, the description of each embodiment has its own emphasis, and for parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0132] Those skilled in the art will appreciate that embodiments of the present invention may provide methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects.

Claims

1. A method for constructing a cardiovascular disease risk prediction model, characterized in that: The method comprises: Get the user's prediction parameters and prediction labels; Constructing a first risk prediction model for predicting the cardiovascular disease risk of the user; Expanding the first risk prediction model to obtain a second risk prediction model; Expanding the second risk prediction model to obtain a third risk prediction model; Using the prediction parameters and the prediction labels, training the first risk prediction model, the second risk model, and the third risk prediction model to obtain a trained first risk prediction model, a trained second risk model, and a trained third risk prediction model; A target risk prediction model for predicting the risk of cardiovascular disease is constructed based on the trained first risk prediction model, the trained second risk model and the trained third risk prediction model.

2. The method for constructing a cardiovascular disease risk prediction model according to claim 1, characterized in that: The constructing of a first risk prediction model for predicting the cardiovascular disease risk of the user comprises: Obtaining a first weight of each of the prediction parameters; Define the weighted summation of the characteristics of each of the prediction parameters using the first weight to obtain the model construction requirements of the corresponding first risk prediction score, construct the model parameters of the first risk prediction model, and obtain the first risk prediction model.

3. The method for constructing a cardiovascular disease risk prediction model according to claim 1, characterized in that: The step of expanding the first risk prediction model to obtain a second risk prediction model includes: Obtaining a second weight of each prediction parameter; Acquire a first extension item for extending the first risk prediction model according to the second weight and the prediction parameter of the existing type; Define the weighted summation of the characteristics of each of the prediction parameters and the first extension item using the second weight to obtain the model construction requirements of the corresponding second risk prediction score, construct the model parameters of the second risk prediction model, and obtain the second risk prediction model.

4. The method for constructing a cardiovascular disease risk prediction model according to claim 1, characterized in that: The step of expanding the second risk prediction model to obtain a third risk prediction model includes: Obtaining a third weight of each prediction parameter; Obtaining a second extension item for extending the second risk prediction model according to the third weight and the prediction parameter of the existing type; Define the weighted summation of the characteristics of each of the prediction parameters and the second extension item using the third weight to obtain the model construction requirements of the corresponding third risk prediction score, construct the model parameters of the third risk prediction model, and obtain the third risk prediction model.

5. The method for constructing a cardiovascular disease risk prediction model according to claim 1, characterized in that: The step of constructing a target risk prediction model for predicting cardiovascular disease risk based on the trained first risk prediction model, the trained second risk model, and the trained third risk prediction model comprises: Get the fourth weight; Define the fourth weight to perform weighted summation on the output results of the trained first risk prediction model, the trained second risk model and the trained third risk prediction model to obtain the model construction requirements of the corresponding target risk prediction score, construct the model parameters of the target risk prediction model, and obtain the target risk prediction model.

6. A system for constructing a cardiovascular disease risk prediction model, characterized in that: The system comprises: Data acquisition module, used to obtain the user's prediction parameters and prediction labels; A first building module, used to build a first risk prediction model for predicting the cardiovascular disease risk of the user; A second building module is used to expand the first risk prediction model to obtain a second risk prediction model; A third building module is used to expand the second risk prediction model to obtain a third risk prediction model; A model training module, used to train the first risk prediction model, the second risk model and the third risk prediction model using the prediction parameters and the prediction labels to obtain a trained first risk prediction model, a trained second risk model and a trained third risk prediction model; The fourth construction module is used to construct a target risk prediction model for predicting the risk of cardiovascular disease based on the trained first risk prediction model, the trained second risk model and the trained third risk prediction model.

Images