A method and device for predicting disease risk

Abstract

The embodiments of the present application provide a method and device for predicting the risk of illness, wherein a method for predicting the risk of illness may include: receiving key factors, screening and processing the physical examination data of the target user according to the key factors, and obtaining risk data, wherein the risk data includes one or more of blood pressure data, blood lipid data, and blood sugar data; determining the risk score of the target user through a target scoring model according to the risk data, wherein the target scoring model is a prediction model trained based on the Xgboost algorithm and the K-Folds cross-validation algorithm method; determining the risk level of the target user's risk of illness according to the risk score. The implementation of the embodiments of the present application can effectively predict the risk of illness of the user, and further, can timely and effectively identify potential threats to the disease in the body, and conduct early screening and intervention for the disease as soon as possible, so as to delay the manifestation of the disease and prevent the aggravation of chronic diseases.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a method and device for predicting disease risk. Background Art

[0002] Cardiovascular disease is the primary disease burden on human health worldwide. The morbidity and mortality of cardiovascular diseases in my country are rising rapidly. It is estimated that the three most common cardiovascular diseases, namely, hypertension, hyperlipidemia and hyperglycemia, can also be said to be the "three highs" diseases, are the culprits of most other cardiovascular and cerebrovascular diseases. Therefore, the prevention and treatment of the "three highs" diseases has become an important need to prevent cardiovascular diseases and protect physical health. At present, the "three highs" prevention examinations in hospitals are obtained based on the specific data of traditional measurements and the professional knowledge of doctors. However, for patients with "three highs" diseases, when only one influencing factor exceeds the standard, the characteristics of the "three highs" diseases may not be shown in time, resulting in people's unclear awareness of the risk of illness and failure to prevent and treat in time.

[0003] Therefore, it is too late to diagnose and prevent diseases based on the specific data of traditional measurements and the professional knowledge of doctors. Furthermore, once the disease is treated, it will cause great losses to the user's health and economic costs. Therefore, how to timely and effectively identify the potential threats that cause diseases in the body, and conduct early screening and intervention for the causes of the disease as soon as possible, so as to delay the manifestation of the disease and prevent chronic diseases from becoming severe. Summary of the invention

[0004] In view of the above problems, the present application is proposed to provide a method and device for predicting disease risk that overcomes the above problems or at least partially solves the above problems.

[0005] In a first aspect, the present application provides a method for predicting disease risk, which may include:

[0006] Receiving a key factor, wherein the key factor is used to indicate a disease type of a target disease or to indicate an influencing factor affecting the target disease;

[0007] Screening and processing the physical examination data of the target user according to the key factors to obtain risk data, wherein the risk data includes one or more of blood pressure data, blood lipid data and blood sugar data;

[0008] Determine the risk score of the target user through a target scoring model according to the risk data, wherein the target scoring model is a prediction model trained based on an Xgboost algorithm and a K-Folds cross-validation algorithm method;

[0009] The risk level of the disease risk of the target user is determined according to the risk score.

[0010] Through the method described in the first aspect, key factors can be received, and the physical examination data of the target user can be screened and processed according to the key factors to obtain one or more of blood pressure data, blood lipid data and blood sugar data; according to the processed data, the risk score corresponding to the target user is determined by the target scoring model trained based on the Xgboost algorithm and the K-Folds cross-validation algorithm method; the risk level of the target user's disease risk is determined according to the risk score. After one or more of the processed blood pressure data, blood lipid data and blood sugar data are input into the risk prediction model, the risk prediction model can effectively predict the risk of the user's disease, and further, it can timely and effectively identify the potential threats of chronic diseases (such as hypertension, hyperglycemia, hyperlipidemia, cardiovascular and cerebrovascular diseases, etc.) caused by the "three highs" factors in the body, and conduct early screening and intervention for the factors causing this type of disease as soon as possible, so as to delay the manifestation of the disease and prevent the severity of chronic diseases.

[0011] In one possible implementation, the physical examination data includes one or more of basic surgical examination data, basic internal medicine examination data, oral examination data, ultrasound data, electrocardiogram data, ophthalmological examination data, gynecological examination data, biochemical examination data, blood pressure measurement results, blood lipid measurement results, and blood sugar measurement results.

[0012] In one possible implementation, the key factors include M keywords, and the M keywords are used to indicate the disease type of the target disease, wherein M is a positive integer greater than or equal to 1; the data screening and processing of the physical examination data of the target user according to the key factors to obtain risk data includes: data screening of the physical examination data according to the M keywords, and determining that the screened physical examination data corresponding to the M keywords are the risk data.

[0013] In one possible implementation, the key factors include M keywords, and the M keywords are used to indicate the disease type of the target disease, wherein M is a positive integer greater than or equal to 1; the physical examination data of the target user is screened and processed according to the key factors to obtain risk data, including: processing the physical examination data corresponding to the M keywords into 0-1 feature data according to a regular expression to obtain the risk data.

[0014] In one possible implementation, before determining the risk score of the target user through the target scoring model based on the risk data, the method also includes: obtaining physical examination data of N users as sample physical examination data, wherein the N users do not include the target user, and N is a positive integer greater than 1; dividing the obtained N sample physical examination data into K-1 training sets and 1 test set according to the K-Folds cross-validation algorithm, wherein K is a positive integer greater than 1; based on the Xgboost algorithm, establishing i regression tree training prediction models on the K-1 training sets, and the prediction models are used to score and predict the risk size of the disease, wherein i is a positive integer greater than 1; according to the 1 test set test, the risk size of the disease is scored and predicted by the trained prediction model; if the result of the score prediction is within a preset error range, the prediction model is determined to be the target scoring model.

[0015] In one possible implementation, the method also includes: based on the Xgboost algorithm, establishing i regression trees on the K training sets, determining s different key factors in the physical examination data that affect the risk data, wherein s is a positive integer greater than 1; the key factors include the s different key factors; the data screening and processing of the physical examination data of the target user according to the key factors, the data screening and processing of the physical examination data of the target user according to the key factors, to obtain the risk data, including: determining the physical examination data of the target user corresponding to the top j key factors with the largest influence values ​​among the influence values ​​of the s key factors, wherein j is a positive integer greater than 1 and less than s; the data screening of the physical examination data of the target user corresponding to the top j key factors, to obtain the risk data.

[0016] In one possible implementation, the method further includes: determining risk scores corresponding to the N sample physical examination data based on the N sample physical examination data and the target scoring model; and dividing risk levels and determining a mapping relationship between the risk scores and the risk levels based on the risk scores corresponding to the N sample physical examination data.

[0017] In a second aspect, an embodiment of the present application provides a device for predicting disease risk, which may include:

[0018] A receiving unit, used to receive a key factor, wherein the key factor is used to indicate a disease type of a target disease or an influencing factor affecting the target disease;

[0019] A processing unit, configured to screen and process the physical examination data of the target user according to the key factors to obtain risk data, wherein the risk data includes one or more of blood pressure data, blood lipid data and blood sugar data;

[0020] A first determination unit is used to determine the risk score of the target user through a target scoring model according to the risk data, wherein the target scoring model is a prediction model trained based on an Xgboost algorithm and a K-Folds cross-validation algorithm device;

[0021] The second determining unit is configured to determine a risk level of the disease risk of the target user according to the risk score.

[0022] In one possible implementation, the physical examination data includes one or more of basic surgical examination data, basic internal medicine examination data, oral examination data, ultrasound data, electrocardiogram data, ophthalmological examination data, gynecological examination data, biochemical examination data, blood pressure measurement results, blood lipid measurement results, and blood sugar measurement results.

[0023] In one possible implementation, the key factors include M keywords, and the M keywords are used to indicate the disease type of the target disease, wherein M is a positive integer greater than or equal to 1; the processing unit is specifically used to: perform data screening on the physical examination data according to the M keywords, and determine that the screened physical examination data corresponding to the M keywords is the risk data.

[0024] In one possible implementation, the key factors include M keywords, and the M keywords are used to indicate the disease type of the target disease, wherein M is a positive integer greater than or equal to 1; the processing unit is specifically used to: process the physical examination data corresponding to the M keywords into 0-1 feature data according to a regular expression method to obtain the risk data.

[0025] In one possible implementation, the device also includes: a model training unit, which is used to obtain physical examination data of N users as sample physical examination data before determining the risk score of the target user through a target scoring model based on the risk data, wherein the N users do not include the target user, and N is a positive integer greater than 1; according to the K-Folds cross-validation algorithm, the obtained N sample physical examination data are divided into K-1 training sets and 1 test set, wherein K is a positive integer greater than 1; based on the Xgboost algorithm, i regression tree training prediction models are established on the K-1 training sets, and the prediction model is used to score and predict the risk size of the disease, wherein i is a positive integer greater than 1; according to the 1 test set test, the risk size of the disease is scored and predicted by the trained prediction model; if the result of the score prediction is within a preset error range, the prediction model is determined to be the target scoring model.

[0026] In one possible implementation, the device also includes: a third determination unit, used to establish i regression trees on the K training sets based on the Xgboost algorithm, and determine s different key factors in the physical examination data that affect the risk data, wherein s is a positive integer greater than 1; the key factors include the s different key factors; the processing unit is specifically used to: determine the physical examination data of the target user corresponding to the top j key factors with the largest influence values ​​among the influence values ​​of the s key factors, wherein j is a positive integer greater than 1 and less than s; and perform data screening on the physical examination data of the target user corresponding to the top j key factors to obtain the risk data.

[0027] In one possible implementation, the device also includes: a fourth determination unit, used to determine the risk scores corresponding to the N sample physical examination data based on the N sample physical examination data and the target scoring model; based on the risk scores corresponding to the N sample physical examination data, divide the risk levels and determine the mapping relationship between the risk scores and the risk levels.

[0028] In a third aspect, an embodiment of the present application provides a device for predicting disease risk, including a storage component, a processing component and a communication component, wherein the storage component is used to store computer programs, and the communication component is used to interact with external devices; the processing component is configured to call the computer program to execute the method described in the first aspect, which will not be repeated here.

[0029] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program, and the computer program is executed by a processor to implement the method of the first aspect above. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the background technology, the drawings required for use in the embodiments of the present application or the background technology will be described below.

[0031] Figure 1 is a schematic diagram of a disease risk prediction system architecture provided in an embodiment of the present application;

[0032] Figure 2A It is a schematic diagram of a process of a method for predicting disease risk provided in an embodiment of the present application;

[0033] Figure 2B This is a schematic diagram of an application interface of a terminal device provided in an embodiment of the present application;

[0034] Figure 3It is a schematic diagram of a process of a method for predicting disease risk provided in an embodiment of the present application;

[0035] Figure 4 It is a schematic diagram of the structure of a device for predicting disease risk provided in an embodiment of the present application;

[0036] Figure 5 It is a schematic diagram of the structure of another device for predicting disease risk provided in an embodiment of the present application. DETAILED DESCRIPTION

[0037] The embodiments of the present application will be described below in conjunction with the drawings in the embodiments of the present application.

[0038] The terms "first", "second", "third", etc. in the specification and claims of the present application and the drawings are used to distinguish different objects, rather than to describe a specific order. In addition, "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally includes steps or units that are not listed, or optionally includes other steps or units inherent to these processes, methods, products or devices.

[0039] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0040] The terms "server", "unit", "system", etc. used in this application are used to refer to computer-related entities, hardware, firmware, a combination of hardware and software, software, or software in execution. For example, a server can be, but is not limited to, a processor, a data processing platform, a computing device, a computer, two or more computers, etc.

[0041] First, some terms in this application are explained to facilitate understanding by those skilled in the art.

[0042] (1) XGBoost algorithm, the full name is XtremeGradient Boosting, where regression tree (CART) is the most basic component of XGBoost. Regression tree refers to the construction of a classification tree based on training features and training data to determine the prediction result of each data. Among them, the Gini index is needed to calculate the gain when building the tree, that is, to select features for building the tree.

[0043] (2) K-Folds cross-validation algorithm, also known as cyclic validation. It divides the original data into K groups (K-Fold), uses each subset of data as a validation set, and uses the remaining K-1 groups of subset data as training sets, so that K models are obtained. The K models are evaluated in the validation set, and the final error MSE (Mean Squared Error) is added and averaged to obtain the cross-validation error. Cross-validation effectively utilizes limited data, and the evaluation results can be as close as possible to the performance of the model on the test set, which can be used as an indicator for model optimization.

[0044] (3) “Three high” diseases refer to high blood pressure, high blood sugar (diabetes) and hyperlipidemia. They are “rich diseases” derived from modern society and may exist alone or be interrelated. For example, diabetics are prone to high blood pressure or hyperlipidemia at the same time, and hyperlipidemia is the main factor in the formation and development of arteriosclerosis. The poor elasticity of blood vessels in patients with arteriosclerosis exacerbates the increase in blood pressure. Therefore, if any of these three diseases occurs, it is easy to develop “three high” diseases in the later stage.

[0045] Next, a prediction system architecture of a disease risk based on which the present application is implemented is described. Figure 1 , Figure 1 It is a schematic diagram of a disease risk prediction system architecture provided in an embodiment of the present application, including: a disease risk prediction device 101 and a terminal device 102.

[0046] The device 101 for predicting the risk of illness may include but is not limited to a background server, a component server, a data processing server, etc. When the device 101 for predicting the risk of illness is a server, the server may communicate with multiple terminals through the Internet, receive physical examination data from terminal devices, and provide terminal users with services for predicting the risk of illness. For example, the server may receive key factors, screen and process the physical examination data of the target user according to the key factors, and obtain risk data, wherein the risk data includes one or more of blood pressure data, blood lipid data, and blood sugar data; determine the risk score of the target user according to the risk data through a target scoring model, wherein the target scoring model is a prediction model trained based on the Xgboost algorithm and the K-Folds cross-validation algorithm method; determine the risk level of the target user's risk of illness according to the risk score.

[0047] The terminal device 102 can install and run related applications. An application refers to a program that corresponds to a server and provides local services to customers. Here, the local service may include but is not limited to: sending physical examination data to the server and receiving prediction reports sent by the server (for example: the prediction report may include the user's risk score and risk level) and other shared information, etc. The terminal in the embodiment of this scheme may include but is not limited to any electronic product based on an intelligent operating system, which can interact with the user through input devices such as keyboards, virtual keyboards, touch pads, touch screens, and voice control devices for human-computer interaction, such as smart phones, tablet computers, personal computers, etc. Among them, intelligent operating systems include but are not limited to any operating system that enriches device functions by providing various mobile applications to mobile devices, such as: Android (Android TM ), iOS TM 、Windows Phone TM wait.

[0048] It is also understandable that Figure 1 The system architecture for predicting disease risk is only a partial exemplary implementation in the embodiments of the present application. The system architecture for predicting disease risk in the embodiments of the present application includes but is not limited to the above system architecture for predicting disease risk.

[0049] Reference Figure 2A , Figure 2A Schematic diagram of a method for predicting disease risk provided in an embodiment of the present application. Figure 1 The system in the following will be combined Figure 2A The description is made from the perspective of the disease risk prediction device 101. The method may include the following steps S201 to S203.

[0050] Step S201: receiving key factors, screening and processing the physical examination data of the target user according to the key factors, and obtaining risk data.

[0051] Specifically, the disease risk prediction device 101 can receive key factors, and screen and process the physical examination data of the target user according to the key factors to obtain risk data, wherein the key factors are used to indicate the disease type of the target disease or indicate the influencing factors affecting the target disease, for example: the key factors may include keywords or key factors, and the risk data include one or more of blood pressure data, blood lipid data, and blood sugar data. It can be understood that the influencing factors affecting the target disease included in the key factors are key factors that can be determined in the physical examination data to affect certain risk data affecting the target disease, and can also be influencing factors that affect the target disease in the physical examination data. Therefore, the key factors mainly indicate factors that can affect the target disease, or indicate the disease type of the target disease. Further, the factors affecting the target disease can be screened out by the disease type. It is also understandable that because the physical examination data is diverse (for example, in text form, picture form, curve form, etc.), it cannot be directly input into the target scoring model; and because the physical examination data is often comprehensive physical examination data, if a certain type of disease needs to be estimated, the physical examination data related to the disease type needs to be screened out, and then the risk of disease is predicted for the screened physical examination data, so that the risk score obtained will be more accurate. Therefore, the implementation of this application requires data processing of the physical examination data, and then the processed data is input into the target scoring model.

[0052] Optionally, the physical examination data may include one or more of basic surgical examination data, basic internal medicine examination data, oral examination data, ultrasound data, electrocardiogram data, ophthalmological examination data, gynecological examination data, biochemical examination data, blood pressure measurement results, blood lipid measurement results, and blood sugar measurement results.

[0053] Step S202: Determine the risk score of the target user through a target scoring model based on the risk data.

[0054] Specifically, the disease risk prediction device 101 can determine the risk score corresponding to the target user according to the disease risk data through a target scoring model, wherein the target scoring model is a prediction model trained based on the Xgboost algorithm and the K-Folds cross-validation algorithm method.

[0055] Optionally, the key factors include M keywords, and the M keywords are used to indicate the disease type of the target disease, wherein M is a positive integer greater than or equal to 1; therefore, the M keywords are received, and the physical examination data is screened according to the M keywords, and the screened physical examination data corresponding to the M keywords are determined to be the risk data. For example, if the keywords are blood pressure, blood sugar or blood lipids, then the related characteristic data such as ophthalmology and oral cavity in the physical examination data can be deleted, while the physical examination data on blood pressure measurement results, blood lipid measurement results and blood sugar measurement results can be retained.

[0056] Optionally, the key factors include M keywords, and the M keywords are used to indicate the disease type of the target disease, wherein M is a positive integer greater than or equal to 1; therefore, M keywords are received, and the physical examination data corresponding to the M keywords are processed into 0-1 feature data according to regular expressions to obtain the risk data. For example: if the keywords are blood pressure, blood sugar or blood lipids, then in the physical examination data, some words such as "fatty liver" and "hyperglycemia" are written in a regular way to form 0-1 features, that is, high blood sugar is considered to be 1, and the numerical features of blood sugar measurement can remain unchanged; or, the words "blood pressure measurement results are normal" in the physical examination data can also be written in a regular way to form 0-1 features, that is, the normal blood pressure measurement result is 0.

[0057] Step S203: Determine the risk level of the target user's disease risk according to the risk score.

[0058] Specifically, the disease risk prediction device 101 can determine the risk level of the disease risk of the target user according to the risk score.

[0059] Optionally, the disease risk prediction device 101 determines a disease risk prediction report according to the risk score, and the disease risk prediction report includes the risk level of the disease risk of the target user corresponding to the risk score. Figure 2B , Figure 2B It is a schematic diagram of an application interface of a terminal device provided in an embodiment of the present application. The target user can use the terminal device 102 to predict the disease risk of the acquired physical examination report to obtain the disease risk prediction report of the user, and timely understand the level of the disease that he may suffer from, so as to seek medical treatment and prevention early. Among them, the disease risk prediction report can also include the preventive measures or treatment means required by the user at this stage corresponding to the risk level. For example: after determining the risk score based on the target scoring model according to the user's physical examination data, the disease risk level can be determined according to the risk score, and finally a disease risk prediction report is generated. According to the report, the preventive measures or treatment means required by each user at the stage are obtained, especially for some potential patients, some potential threat factors found according to the above algorithm are targeted for prevention and intervention.

[0060] By implementing the embodiments of the present application, key factors can be received, and the physical examination data of the target user can be screened and processed according to the key factors to obtain one or more of blood pressure data, blood lipid data, and blood sugar data; according to the processed data, the risk score corresponding to the target user is determined by the target scoring model trained based on the Xgboost algorithm and the K-Folds cross-validation algorithm method; the risk level of the target user's disease risk is determined according to the risk score. After inputting one or more of the processed blood pressure data, blood lipid data, and blood sugar data into the risk prediction model, the risk prediction model can effectively predict the risk of the user's disease. Further, it can timely and effectively identify the potential threats of chronic diseases (such as hypertension, hyperglycemia, hyperlipidemia, cardiovascular and cerebrovascular diseases, etc.) caused by the "three highs" factors in the body, and conduct early screening and intervention for the factors that cause this type of disease as soon as possible, so as to delay the manifestation of the disease and prevent the severity of chronic diseases.

[0061] Reference Figure 3 , Figure 3 Schematic diagram of a method for predicting disease risk provided in an embodiment of the present application. Figure 1 The system in the following will be combined Figure 3 The description is made from the perspective of the disease risk prediction device 101. The method may include the following steps S301 to S308.

[0062] Step S301: Acquire physical examination data of N users as sample physical examination data.

[0063] Specifically, the disease risk prediction device 101 can obtain the physical examination data of N users from a physical examination center or other data platform as sample physical examination data. Because the physical examination data of N users are used as sample data for training the target scoring model, the physical examination data of the N users do not include the physical examination data of the target user, and N is a positive integer greater than 1.

[0064] Step S302: Divide the obtained N sample physical examination data into K-1 training sets and 1 test set according to the K-Folds cross-validation algorithm.

[0065] Specifically, the device 101 for predicting the risk of disease divides the obtained N sample physical examination data into K-1 training sets and 1 test set according to the K-Folds cross-validation algorithm, where K is a positive integer greater than 1. K training sets and test sets are obtained from the N sample physical examination data through the K-Folds cross-validation algorithm: the data set is divided into K parts, so that K-1 parts are used to train the model each time, and the remaining 1 part of the data is used to verify the model effect.

[0066] Step S303: Based on the Xgboost algorithm, i regression tree training prediction models are established on the K-1 training sets.

[0067] Specifically, the prediction device 101 for disease risk establishes i regression tree training prediction models on the K-1 training sets based on the Xgboost algorithm, and the prediction models are used to score and predict the risk of disease, where i is a positive integer greater than 1. That is, m regression trees are established on the training set for prediction: each regression tree is a weak learner, and by adding new weak learners step by step, the residuals of all previous weak learners are corrected. Therefore, the final addition of these m learners can be used to predict the final disease risk.

[0068] Optionally, in the process of training the model based on the Xgboost algorithm by the prediction device 101 for disease risk, the objective function of the training process can be optimized: the predicted value of the tree group is as close to the true value (ie, accuracy) as possible and has the greatest possible generalization ability. It should be noted that the prediction error in the objective function must be as small as possible, that is, the leaf nodes of the tree should be as few as possible and the node values ​​should be distributed as little as possible. In the optimization process of the Xgboost algorithm, the eigenvalue that minimizes its loss function is found and split step by step. The splitting node of each eigenvalue is determined by the node that minimizes the root mean square error of the eigenvalue.

[0069] Step S304: Based on a test set test, the risk of illness is scored and predicted using the trained prediction model.

[0070] Specifically, the device 101 for predicting the risk of illness is tested according to one test set, and the risk of illness is scored and predicted by the trained prediction model. That is, after the prediction model is trained by K-1 training sets, the remaining one test set data needs to be substituted into the trained prediction model to perform a score prediction on the risk of illness. If the result of the score prediction is not within the preset error range, the prediction model cannot accurately predict the user's risk of illness, and the prediction model needs further training.

[0071] Step S305: If the result of the rating prediction is within a preset error range, the prediction model is determined to be a target rating model.

[0072] Specifically, when the disease risk prediction device 101 substitutes the remaining test set data into the trained prediction model to perform a scoring prediction on the disease risk, if the result of the scoring prediction is within a preset error range, the prediction model is determined to be the target scoring model, proving that the prediction model can accurately predict the user's disease risk.

[0073] Optionally, the device 101 for predicting the risk of disease can also determine the risk scores corresponding to the N sample physical examination data based on the N sample physical examination data and the target scoring model, and divide the risk levels and determine the mapping relationship between the risk scores and the risk levels based on the risk scores corresponding to the N sample physical examination data. It can be understood that the higher the risk score, the higher the corresponding risk level, and the greater the probability that the user may suffer from the disease. Therefore, based on this mapping relationship, the risk level corresponding to the risk score of the target user can be determined, and then the probability of the target user being ill can be judged.

[0074] Step S306: Screen and process the physical examination data of the target user according to key factors to obtain risk data.

[0075] Specifically, the prediction device 101 of the disease risk is based on the Xgboost algorithm, and i regression trees are established on the K training sets to determine s different key factors in the physical examination data that affect the risk data, where s is a positive integer greater than 1. The key factors include the s different key factors; the physical examination data of the target user is screened and processed according to the key factors to obtain the risk data, including: determining the physical examination data of the target user corresponding to the first j key factors with the largest influence value among the influence values ​​of the s key factors, where j is a positive integer greater than 1 and less than s; the physical examination data of the target user corresponding to the first j key factors are screened to obtain the risk data. It can be understood that the physical examination data of the target user corresponding to the first j key factors can be screened out from the physical examination data of the target user, and then the final risk data can be obtained after data processing, and the risk data can be processed into 0-1 feature data according to the regular expression method or the text data can be converted into numerical data.

[0076] Step S307: Determine the risk score of the target user according to the risk data through a target scoring model.

[0077] Step S308: Determine the risk level of the target user's disease risk according to the risk score.

[0078] Specifically, the implementation of step S307-step S308 can also refer to the above Figure 2A The relevant descriptions of steps S202 to S203 are not repeated here.

[0079] In implementing the embodiment of the present application, the user's physical examination data can be first processed so that the physical examination data is in numerical form for calculation, and then the processed user data is input into the risk prediction model, and the risk prediction model predicts the risk of the user's illness. Among them, the risk prediction model is based on a large amount of sample data using univariate analysis correlation or decision tree to determine the key factors affecting the Y feature data in the sample data, and then the target key factors and the K training sets and test sets obtained according to the K-Folds cross-validation algorithm and N sample physical examination data are input into the prediction model trained in the "three highs" disease risk prediction model established by the xgboost algorithm. Secondly, the trained risk scoring model can be used to determine the risk of the target user's illness, and the internal laws of some "three highs" diseases can be discovered, the influencing factors of the disease can be controlled, the formation of the disease can be prevented, or the disease can be inhibited from developing, to provide AI support for solving the national medical insurance cost control problem and improving the level of medical and health services.

[0080] The above describes in detail the method of the embodiment of the present application. The following provides a prediction device for disease risk related to the embodiment of the present application. The prediction device 10 for disease risk can be a service device that brings various conveniences to third parties by acquiring, processing, analyzing and extracting valuable data. Please refer to the attached Figure 4 , Figure 4 It is a structural diagram of a device for predicting disease risk provided by an embodiment of the present application, which may include a receiving unit 401, a processing unit 402, a first determining unit 403, a second determining unit 404, and may also include: a model training unit 405, a third determining unit 406, and a fourth determining unit 407.

[0081] A receiving unit 401 is used to receive a key factor, where the key factor is used to indicate a disease type of a target disease or an influencing factor affecting the target disease;

[0082] The processing unit 402 is used to screen and process the physical examination data of the target user according to the key factors to obtain risk data, where the risk data includes one or more of blood pressure data, blood lipid data and blood sugar data;

[0083] A first determination unit 403 is used to determine the risk score of the target user through a target scoring model according to the risk data, wherein the target scoring model is a prediction model trained based on an Xgboost algorithm and a K-Folds cross-validation algorithm device;

[0084] The second determining unit 404 is configured to determine the risk level of the disease risk of the target user according to the risk score.

[0085] In one possible implementation, the physical examination data includes one or more of basic surgical examination data, basic internal medicine examination data, oral examination data, ultrasound data, electrocardiogram data, ophthalmological examination data, gynecological examination data, biochemical examination data, blood pressure measurement results, blood lipid measurement results, and blood sugar measurement results.

[0086] In one possible implementation, the key factors include M keywords, and the M keywords are used to indicate the disease type of the target disease, wherein M is a positive integer greater than or equal to 1; the processing unit 402 is specifically used to: filter the physical examination data according to the M keywords, and determine that the filtered physical examination data corresponding to the M keywords are the risk data.

[0087] In one possible implementation, the key factors include M keywords, and the M keywords are used to indicate the disease type of the target disease, where M is a positive integer greater than or equal to 1; the processing unit 402 is specifically used to: process the physical examination data corresponding to the M keywords into 0-1 feature data according to a regular expression method to obtain the risk data.

[0088] In one possible implementation, the device also includes: a model training unit 405, which is used to obtain physical examination data of N users as sample physical examination data before determining the risk score of the target user through a target scoring model based on the risk data, wherein the N users do not include the target user, and N is a positive integer greater than 1; according to the K-Folds cross-validation algorithm, the obtained N sample physical examination data are divided into K-1 training sets and 1 test set, wherein K is a positive integer greater than 1; based on the Xgboost algorithm, i regression tree training prediction models are established on the K-1 training sets, and the prediction model is used to score and predict the risk size of the disease, wherein i is a positive integer greater than 1; according to the 1 test set test, the risk size of the disease is scored and predicted by the trained prediction model; if the result of the score prediction is within a preset error range, the prediction model is determined to be the target scoring model.

[0089] In one possible implementation, the device also includes: a third determination unit 406, which is used to establish i regression trees on the K training sets based on the Xgboost algorithm, and determine s different key factors in the physical examination data that affect the risk data, wherein s is a positive integer greater than 1; the key factors include the s different key factors; the processing unit 402 is specifically used to: determine the physical examination data of the target user corresponding to the top j key factors with the largest influence values ​​among the influence values ​​of the s key factors, wherein j is a positive integer greater than 1 and less than s; and perform data screening on the physical examination data of the target user corresponding to the top j key factors to obtain the risk data.

[0090] In one possible implementation, the device also includes: a fourth determination unit 407, which is used to determine the risk scores corresponding to the N sample physical examination data based on the N sample physical examination data and the target scoring model; based on the risk scores corresponding to the N sample physical examination data, divide the risk levels and determine the mapping relationship between the risk scores and the risk levels.

[0091] It should be noted that the implementation of each operation can also refer to Figure 2A-Figure 3 The corresponding description of the method embodiment shown will not be repeated here.

[0092] like Figure 5 As shown, Figure 5 2 is a schematic diagram of the structure of another device for predicting disease risk provided in an embodiment of the present application, wherein the device 20 includes at least one processor 501, at least one memory 502, and at least one communication interface 503. In addition, the device may also include common components such as an antenna, which will not be described in detail here.

[0093] The processor 501 may be a general purpose central processing unit (CPU), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits for controlling the execution of the above program.

[0094] The communication interface 503 is used to communicate with other devices or communication networks, such as Ethernet, Radio Access Network (RAN), core network, Wireless Local Area Networks (WLAN), etc.

[0095] The memory 502 may be a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM) or other types of dynamic storage devices that can store information and instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical disc, laser disc, optical disc, digital versatile disc, Blu-ray disc, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto. The memory may exist independently and be connected to the processor via a bus. The memory may also be integrated with the processor.

[0096] The memory 502 is used to store application code for executing the above solution, and the execution is controlled by the processor 501. The processor 501 is used to execute the application code stored in the memory 502.

[0097] The code stored in the memory 502 can execute the above Figure 2A-Figure 3 The provided disease risk prediction method, for example, when the device 20 is predicting disease risk 101, can receive key factors, screen and process the physical examination data of the target user according to the key factors, and obtain risk data, wherein the risk data includes one or more of blood pressure data, blood lipid data, and blood sugar data; determine the risk score of the target user according to the risk data through a target scoring model, and the target scoring model is a prediction model trained based on the Xgboost algorithm and the K-Folds cross-validation algorithm method; determine the risk level of the disease risk of the target user according to the risk score.

[0098] It should be noted that the functions of the various functional units in the disease risk prediction device 20 described in the embodiment of the present application can refer to Figure 2A-Figure 3 The corresponding description of the method embodiment shown will not be repeated here.

[0099] In the present application, the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the embodiments of the present application.

[0100] In addition, the functional components in the various embodiments of the present application may be integrated into one component, or each component may exist physically separately, or two or more components may be integrated into one component. The above integrated components may be implemented in the form of hardware or in the form of software functional units.

[0101] If the integrated components are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or partly contributed to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0102] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed in the present application, and these modifications or replacements should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

[0103] It should be understood that in various embodiments of the present application, the size of the sequence number of each process does not mean the order of execution, and the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application. Although the present application is described herein in conjunction with various embodiments, however, in the process of the present application claimed for protection by the embodiments, those skilled in the art can understand and implement other changes of the disclosed embodiments.

Claims

1. A method for predicting disease risk, characterized in that: include: Acquire physical examination data of N users as sample physical examination data; wherein the N users do not include the target user, and N is a positive integer greater than 1; According to the K-Folds cross-validation algorithm, the N sample physical examination data obtained are divided into K-1 training sets and 1 test set; Based on the Xgboost algorithm, i regression tree training prediction models are established on the K-1 training sets; wherein i is a positive integer greater than 1; According to a test set test, the risk of illness is scored and predicted by the trained prediction model; If the result of the score prediction is within the preset error range, the prediction model is determined to be the target score model; Based on the Xgboost algorithm, i regression trees are established on the K-1 training sets to determine s different key factors affecting the risk data in the physical examination data, where s is a positive integer greater than 1; receiving a key factor, wherein the key factor is used to indicate an influencing factor affecting a target disease, and the key factor includes the s different key factors; The physical examination data of the target user is screened and processed according to the s different key factors to obtain risk data, including: determining the physical examination data of the target user corresponding to the first j key factors with the largest influence values ​​among the influence values ​​of the s different key factors, wherein j is a positive integer greater than 1 and less than s, and the physical examination data of the target user corresponding to the first j key factors is screened to obtain the risk data, wherein the risk data includes one or more of blood pressure data, blood lipid data, and blood sugar data; Determine the risk score of the target user through the target scoring model according to the risk data; The risk level of the disease risk of the target user is determined according to the risk score.

2. The method according to claim 1, characterized in that: The physical examination data includes basic surgical examination data and basic internal medicine examination data.

3. The method according to claim 1, characterized in that: The method further comprises: Determining risk scores corresponding to the N sample physical examination data according to the N sample physical examination data and the target scoring model; According to the risk scores corresponding to the N sample physical examination data, risk levels are divided and a mapping relationship between the risk scores and the risk levels is determined.

4. A device for predicting disease risk, characterized in that: include: A model training unit is used to obtain the physical examination data of N users as sample physical examination data, wherein the N users do not include the target user, and N is a positive integer greater than 1; according to the K-Folds cross-validation algorithm, the obtained N sample physical examination data are divided into K-1 training sets and 1 test set, wherein K is a positive integer greater than 1; based on the Xgboost algorithm, i regression tree training prediction models are established on the K-1 training sets, wherein i is a positive integer greater than 1; according to the 1 test set test, the risk of illness is scored and predicted by the trained prediction model; if the result of the score prediction is within a preset error range, the prediction model is determined to be a target scoring model; A third determination unit is used to establish i regression trees on the K-1 training sets based on the Xgboost algorithm to determine s different key factors affecting the risk data in the physical examination data, where s is a positive integer greater than 1; A receiving unit, used to receive a key factor, where the key factor is used to indicate an influencing factor affecting a target disease, and the key factor includes s different key factors; a processing unit, configured to screen and process the physical examination data of the target user according to the s different key factors to obtain risk data, wherein the risk data includes one or more of blood pressure data, blood lipid data and blood sugar data; The processing unit is specifically configured to determine the physical examination data of the target user corresponding to the first j key factors with the largest influence values ​​among the influence values ​​of the s different key factors, wherein j is a positive integer greater than 1 and less than s, and to perform data screening on the physical examination data of the target user corresponding to the first j key factors to obtain the risk data; A first determining unit, configured to determine a risk score of the target user according to the risk data through the target scoring model; The second determining unit is configured to determine a risk level of the disease risk of the target user according to the risk score.

5. A device for predicting disease risk, characterized in that: The method comprises a processing component, a storage component and a communication module component, wherein the processing component, the storage component and the communication component are interconnected, wherein the storage component is used to store computer programs, and the communication component is used to exchange information with external devices; the processing component is configured to call the computer program to execute the method as described in any one of claims 1-3.

6. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and the computer program is executed by a processor to implement the method according to any one of claims 1 to 3.

Images