Hypertension risk assessment method and device based on historical data, and terminal

By constructing a multi-level deep learning neural network model and combining nonlinear blood pressure interval allocation and multiple health impact weight values, the problem of low accuracy in the risk assessment of hypertension in the existing technology is solved, and more accurate risk assessment and personality of hypertension is achieved. Health management.

CN119092109AInactive Publication Date: 2024-12-06SHENZHEN LIKANG TECHNOLOGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410992272.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-23
Publication Date
2024-12-06
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The prediction accuracy of the existing hypertension risk assessment methods is not high enough, and lacks the characteristics of integrating multiple indicator characteristics, which cannot effectively reflect the differences in hypertension risk in different races and regions.

Method used

Using a hypertension risk assessment method based on historical data, a multi-level deep learning neural network model was constructed to obtain data sets of multiple hypertensive patients for preprocessing, and after grouping, training and cross-validation were performed, combining nonlinear blood pressure interval allocation and multiple to improve prediction accuracy.

Benefits of technology

It improves the prediction accuracy of hypertension risk assessment, can more accurately reflect the hypertension risks of different groups of people, and provides more personalized health management strategies.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119092109A_ABST
    Figure CN119092109A_ABST
Patent Text Reader

Abstract

The invention discloses a hypertension risk assessment method and device based on historical data and a terminal. The method comprises the steps that an initial model is built, wherein the initial model comprises a plurality of hidden layers; an initial data set is obtained, the initial data set is preprocessed, a target training set and a target support set are obtained, and the initial data set comprises data sets of multiple hypertensive patients; training the initial model based on the target training set, and performing cross validation and hyper-parameter adjustment based on the target support set to obtain a target prediction model; target data are obtained, the target data are predicted based on the target prediction model, a hypertension risk prediction result of a target patient is obtained, and the target data are hypertension related data of the target patient. According to the hypertension risk assessment method based on historical data provided by the invention, the hypertension risk of the patient is predicted based on the multi-level network model by fusing multiple index features, the prediction precision is higher, and the prediction result is more accurate.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of computer-aided diagnosis, and in particular to a method, device and terminal for evaluating hypertension risk based on historical data. Background Art

[0002] With the development of social economy, people's living standards are constantly improving, and the incidence of chronic diseases is rising. Chronic diseases are a general term for diseases that are not contagious and have long-term accumulation to form disease-like damage. Common chronic diseases mainly include cardiovascular and cerebrovascular diseases, cancer, diabetes, and chronic respiratory diseases. Among them, cardiovascular and cerebrovascular diseases include hypertension, stroke, and coronary heart disease. Among them, hypertension is a relatively common chronic disease and one of the main risk factors for cardiovascular and cerebrovascular diseases. Long-term uncontrolled hypertension will increase the incidence of heart disease, stroke, kidney disease, etc. The symptoms of hypertension are generally not obvious and are often called "asymptomatic diseases", but long-term hypertension can cause damage to the body, such as headaches, dizziness, blurred vision, tinnitus, palpitations, etc. Therefore, risk prediction for hypertension is particularly important.

[0003] In the existing technology, with the development of machine learning and deep learning, deep learning technology is also used to predict hypertension. However, the existing machine prediction method of hypertension risk is based on a single data and does not integrate the characteristics of multiple indicators, resulting in the problem of insufficient prediction accuracy.

[0004] Therefore, the prior art still needs to be improved and enhanced. Summary of the invention

[0005] In view of the above-mentioned defects of the prior art, a method, device and terminal for evaluating hypertension risk based on historical data are provided, aiming to solve the problem of insufficient prediction accuracy when predicting hypertension risk level in the prior art.

[0006] A first aspect of the present invention provides a method for assessing hypertension risk based on historical data, comprising:

[0007] Constructing an initial model, wherein the initial model includes multiple hidden layers;

[0008] Acquire an initial data set, and preprocess the initial data set to obtain a target training set and a target support set, wherein the initial data set includes data sets of multiple hypertensive patients;

[0009] The initial model is trained based on the target training set, and cross-validation and hyperparameter adjustment are performed based on the target support set to obtain a target prediction model;

[0010] Target data is acquired, and the target data is predicted based on the target prediction model to obtain a hypertension risk prediction result of the target patient, wherein the target data is hypertension-related data of the target patient.

[0011] In one implementation, constructing the initial model includes:

[0012] Based on the first formula, a deep learning neural network with K hidden layers is constructed;

[0013] The first formula is:

[0014]

[0015] Where X = H (0) is the input matrix of the original data, H (k) ,k=1,2,L,K is the k-th layer depth representation of X, W (k) is the weight of the kth layer of the initial model.

[0016] In one implementation, the weight of the kth layer of the initial model is obtained based on the second formula;

[0017] The second formula is:

[0018]

[0019] Where V (k) is a random value matrix.

[0020] In one implementation, the preprocessing of the initial data set includes:

[0021] Converting discrete value attributes in the initial data set into continuous value attributes to obtain a first data set;

[0022] Randomly extract 80% of the data in the first data set as an initial training set, and the remaining 20% ​​of the data in the first data set as an initial support set;

[0023] The blood pressure data in the initial training set and the initial support set are obtained, and the initial training set and the initial support set are grouped based on the blood pressure data to obtain the target training set and the target support set, wherein the blood pressure data are diastolic pressure and systolic pressure data.

[0024] In one implementation, the grouping the initial training set and the initial support set based on the blood pressure data to obtain the target training set and the target support set includes:

[0025] The diastolic pressure and systolic pressure in the initial training set and the initial support set are extracted and divided into multiple target intervals in a nonlinear manner;

[0026] The initial training set and the initial support set are grouped based on the target interval to obtain the target training set and the target support set.

[0027] In one implementation, the training of the initial model based on the target training set includes:

[0028] Constructing an initial interval impact weight value corresponding to each target interval;

[0029] Constructing initial differentiated health impact weight values ​​within each target interval, wherein the differentiated health impact weight values ​​include weight values ​​corresponding to age, gender, race, region, body mass index, family medical history, dietary preferences, exercise frequency, mental health status, and pregnancy status;

[0030] After the initial model is trained based on the target training set, cross-validation and hyperparameter adjustment are performed based on the target support set to obtain the target interval impact weight value and the target differentiated health impact weight value.

[0031] In one implementation, the training of the initial model based on the target training set includes:

[0032] For non-convex logistic regression loss functions, construct an optimization problem with target complementarity constraints;

[0033] The target complementarity constraint optimization problem is:

[0034]

[0035] stg(x)≥0,h(x)≥0,<g(x),h(x)> ≥0;

[0036] The target complementarity constraint optimization problem is solved by a smooth approximation method based on minimizing the entropy function.

[0037] A second aspect of the present invention provides a hypertension risk assessment device based on historical data, comprising:

[0038] A model building module, used to build an initial model, wherein the initial model contains multiple hidden layers;

[0039] A data set acquisition module is used to acquire an initial data set, preprocess the initial data set, and obtain a target training set and a target support set, wherein the initial data set includes data sets of multiple hypertensive patients;

[0040] A training module, used to train the initial model based on the target training set, perform cross-validation and hyperparameter adjustment based on the target support set, and obtain a target prediction model;

[0041] The prediction module is used to obtain target data, predict the target data based on the target prediction model, and obtain the hypertension risk prediction result of the target patient, wherein the target data is hypertension-related data of the target patient.

[0042] The third aspect of the present invention provides a terminal, comprising: a processor, a storage medium communicatively connected to the processor, the storage medium being suitable for storing a plurality of instructions, and the processor being suitable for calling the instructions in the storage medium to execute the steps of implementing any of the above-mentioned methods for assessing hypertension risk based on historical data.

[0043] A fourth aspect of the present invention provides a storage medium, wherein the storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps of any of the above-mentioned methods for assessing hypertension risk based on historical data.

[0044] Beneficial effects: Compared with the prior art, the present invention provides a hypertension risk assessment method, device and terminal based on historical data. In the hypertension risk assessment method based on historical data provided by the present invention, when performing hypertension risk assessment on a target patient, an initial model containing multiple hidden layers is constructed, and then an initial data set including multiple hypertensive patients is obtained, the initial data set is preprocessed to obtain a target training set and a target support set, the initial model is trained based on the target training set, cross-validation and hyperparameter adjustment are performed based on the target support set, and a target prediction model is obtained. Finally, the target data is obtained, and the target data is predicted based on the target prediction model to obtain a hypertension risk prediction result for the target patient, and the target data is hypertension-related data for the target patient. The hypertension risk assessment method based on historical data provided by the present invention uses a personal health score square mean evaluation algorithm with big data, nonlinearity, and medium-granularity blood pressure interval allocation, integrates multiple indicator features, constructs a target prediction model, and predicts the patient's hypertension risk based on a multi-level network model, so that the prediction accuracy is higher and the prediction results are more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 A flowchart of an embodiment of a hypertension risk assessment method based on historical data provided by the present invention;

[0046] Figure 2 A schematic diagram of the structure of an embodiment of a hypertension risk assessment device based on historical data provided by the present invention;

[0047] Figure 3 A structural principle diagram of an embodiment of a terminal provided by the present invention. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical solution and effect of the present invention clearer and more specific, the present invention is further described in detail with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0049] It will be understood by those skilled in the art that, unless expressly stated, the singular forms "one", "said", and "the" used herein may also include plural forms. It should be further understood that the term "comprising" used in the specification of the present invention refers to the presence of the features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof. It should be understood that when we refer to an element as being "connected" or "coupled" to another element, it may be directly connected or coupled to the other element, or there may be intermediate elements. In addition, the "connection" or "coupling" used herein may include wireless connection or wireless coupling. The term "and / or" used herein includes all or any unit and all combinations of one or more associated listed items.

[0050] It will be understood by those skilled in the art that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as generally understood by those skilled in the art in the art to which the present invention belongs. It should also be understood that terms such as those defined in general dictionaries should be understood to have meanings consistent with the meanings in the context of the prior art, and will not be interpreted with idealized or overly formal meanings unless specifically defined as herein.

[0051] The hypertension risk assessment method based on historical data provided by the present invention can be applied to a terminal with computing capabilities, and the terminal can execute the hypertension risk assessment method based on historical data provided by the present invention to detect and locate the target position in the image to be processed.

[0052] Embodiment 1

[0053] At present, with the development of social economy, people's living standards are constantly improving, and the incidence of chronic diseases such as hypertension is constantly rising. Hypertension is one of the main risk factors for cardiovascular and cerebrovascular diseases. Long-term uncontrolled hypertension will increase the incidence of heart disease, stroke, coronary heart disease, etc.; it will cause kidney vascular damage, causing renal insufficiency, glomerular sclerosis and other diseases; in severe cases, it will also cause retinopathy, leading to decreased vision and blindness; it will also cause insufficient blood supply to the brain, causing symptoms such as headaches, dizziness, and insomnia. Furthermore, hypertension is closely related to metabolic diseases such as diabetes and hyperlipidemia, which affect each other and aggravate the harm of the disease. Therefore, risk prediction for hypertension is particularly important.

[0054] In the existing technology, as machine learning and deep learning are widely used in predicting hypertension, feature screening methods based on big data and integrating multiple indicators are also gradually emerging, providing standards for early warning of diseases and accurate staging and classification of diseases, and to a large extent providing clinicians with a basis for auxiliary diagnosis.

[0055] However, existing hypertension risk assessment methods are usually determined by diastolic and systolic blood pressure, which is relatively simple and does not integrate the characteristics of multiple indicators. They cannot reflect the differences in hypertension risks among different races and regions, and cannot predict acute cardiovascular events.

[0056] In response to the above-mentioned defects, this embodiment provides a hypertension risk assessment method based on historical data. Specifically, this embodiment obtains a personal blood pressure health assessment score and performs personal blood pressure health prediction and early warning through a square mean assessment algorithm of personal health scores assigned by big data, nonlinearity, and medium-granularity blood pressure intervals. That is, diastolic pressure and systolic pressure are divided into multiple blood pressure intervals in a nonlinear manner, and a number of doctors make subjective score assessments of the health effects of different blood pressure intervals on different groups of people of different ages and genders. The subjective assessment values ​​of many doctors are squared and averaged (expanded) to obtain the health impact weights of different blood pressure intervals on different groups of people (blood pressure interval health weights). The intelligent physical examination station will conduct big data analysis based on the measured blood pressure values, blood pressure interval weights and other personal health data, make health risk predictions for chronic disease populations, and formulate chronic disease health assessment standards suitable for different populations.

[0057] like Figure 1 As shown, the hypertension risk assessment method based on historical data provided in this embodiment includes the following steps:

[0058] S100: construct an initial model, wherein the initial model contains multiple hidden layers.

[0059] Specifically, this embodiment predicts hypertension by constructing an efficient health risk prediction model that can handle hyperparameters and mixed-value attribute data.

[0060] The initial model constructed contains multiple hidden layers. In this embodiment, a deep learning neural network containing K hidden layers is constructed based on the first formula;

[0061] Among them, the first formula is:

[0062]

[0063] Where X = H (0) is the input matrix of the original data, H (k) ,k=1,2,L,K is the k-th layer depth representation of X, W(k) is the weight of the kth layer of the initial model.

[0064] Then, the weight of the kth layer of the initial model is obtained based on the second formula;

[0065] The second formula is:

[0066]

[0067] Where V (k) is a random value matrix.

[0068] S200, obtaining an initial data set, and preprocessing the initial data set to obtain a target training set and a target support set, wherein the initial data set includes data sets of multiple hypertensive patients.

[0069] Specifically, this embodiment mainly considers the risk assessment of hypertension under ultra-healthy parameters. In view of the defects of existing hypertension prediction methods, when obtaining initial data, hypertension-related data of at least 5,000 patients are obtained through multiple channels. These data cover a series of biomedical and lifestyle indicators, including but not limited to age, gender, weight, body mass index (BMI), smoking and drinking history, blood pressure value, cholesterol level, blood sugar value, exercise frequency, eating habits, and family medical history, etc. Furthermore, in this embodiment, in order to study the differences in hypertension risks among different races and regions, the race, birthplace, and permanent residence of the patient are also obtained. These data are a collection of records from multiple medical institutions, ensuring the diversity and representativeness of the data.

[0070] However, due to the diversity of data, after obtaining the initial data, the initial data needs to be preprocessed to obtain a suitable training set and support set.

[0071] Specifically, the preprocessing of the initial data set includes:

[0072] Converting discrete value attributes in the initial data set into continuous value attributes to obtain a first data set;

[0073] Randomly extract 80% of the data in the first data set as an initial training set, and the remaining 20% ​​of the data in the first data set as an initial support set;

[0074] The blood pressure data in the initial training set and the initial support set are obtained, and the initial training set and the initial support set are grouped based on the blood pressure data to obtain the target training set and the target support set, wherein the blood pressure data are diastolic pressure and systolic pressure data.

[0075] Specifically, in order to solve the problem of reducing the mixed-value attribute super-healthy parameters, in this embodiment, it is first necessary to convert the data of the discrete-value attribute in the initial data set into continuous-value attributes. Specifically, the discrete-value attributes in the initial data set are encoded in a one-hot manner, each category is converted into a binary column, and then a deep neural network is constructed to transform the one-hot encoding to obtain the first data set. In the process of training the neural network, the loss function adopts a combination of mean square error and information volume to ensure that the continuous-value attribute after conversion has the lowest error while minimizing the loss of information volume.

[0076] Next, in order to ensure the randomness of the data, 80% of the data in the first data set is randomly selected as an initial training set, and the remaining 20% ​​of the data in the first data set is used as an initial support set.

[0077] At the same time, in this embodiment, it also includes: obtaining the blood pressure data in the initial training set and the initial support set, grouping the initial training set and the initial support set based on the blood pressure data, and obtaining the target training set and the target support set, wherein the blood pressure data are diastolic pressure and systolic pressure data.

[0078] The grouping the initial training set and the initial support set based on the blood pressure data to obtain the target training set and the target support set includes:

[0079] The diastolic pressure and systolic pressure in the initial training set and the initial support set are extracted and divided into multiple target intervals in a nonlinear manner;

[0080] The initial training set and the initial support set are grouped based on the target interval to obtain the target training set and the target support set.

[0081] That is to say, in this embodiment, the diastolic pressure and systolic pressure in the first data set are divided into multiple blood pressure intervals in a nonlinear manner, and a number of doctors make subjective score assessments of the health effects of different blood pressure intervals on different groups of people of different ages, genders, races, regions, etc. The subjective assessment values ​​of many doctors are squared and averaged to expand the differences, thereby obtaining the weights of the health effects of different blood pressure intervals on different groups of people. Furthermore, the weights corresponding to different ages, genders, races, and regions in each blood pressure interval can also be calculated to reflect the differences in hypertension risks among different races and regions.

[0082] S300, training the initial model based on the target training set, performing cross-validation and hyperparameter adjustment based on the target support set, and obtaining a target prediction model;

[0083] The training of the initial model based on the target training set includes:

[0084] Constructing an initial interval impact weight value corresponding to each target interval;

[0085] Constructing initial differentiated health impact weight values ​​within each target interval, wherein the differentiated health impact weight values ​​include weight values ​​corresponding to age, gender, race, region, body mass index, family medical history, dietary preferences, exercise frequency, mental health status, and pregnancy status;

[0086] After the initial model is trained based on the target training set, cross-validation and hyperparameter adjustment are performed based on the target support set to obtain the target interval impact weight value and the target differentiated health impact weight value.

[0087] Specifically, in this embodiment, the initial model training is to first construct the initial health impact weight value corresponding to each target interval. That is to say, after the diastolic pressure and systolic pressure are divided into multiple blood pressure intervals in a nonlinear manner and multiple target intervals are obtained, the weight value of the target interval is first constructed, and the diagnosis results given by the doctor corresponding to the target interval are used to train the model to obtain the initial interval impact weight value corresponding to each target interval.

[0088] Then, the initial differentiated health impact weight values ​​within each of the target intervals are constructed, and the differentiated health impact weight values ​​include the weight values ​​corresponding to age, gender, race, region, body mass index, family medical history, dietary preferences, exercise frequency, mental health status and pregnancy status. Similarly, within each of the target intervals, the weight values ​​corresponding to age, gender, race, region, body mass index, family medical history, dietary preferences, exercise frequency, mental health status and pregnancy status are constructed, and then training is performed based on the diagnosis results given by the doctor to obtain the target differentiated health impact weight values ​​corresponding to each of the target intervals.

[0089] In this way, when the model makes the final prediction, big data analysis can be performed based on the measured blood pressure value, the target interval weight, and the differentiated health impact weight value to obtain a prediction result based on comprehensive considerations of multiple aspects.

[0090] The training of the initial model based on the target training set includes:

[0091] For non-convex logistic regression loss functions, construct an optimization problem with target complementarity constraints;

[0092] The target complementarity constraint optimization problem is:

[0093]

[0094] stg(x)≥0,h(x)≥0,<g(x),h(x)> ≥0;

[0095] The target complementarity constraint optimization problem is solved by a smooth approximation method based on minimizing the entropy function.

[0096] Specifically, when calculating the initial health impact weight value and the differentiated health impact weight value, it involves solving the non-convex logistic regression loss function. At this time, the target complementary constraint optimization problem is constructed and then solved. Among them, the use of a non-negative vertical constraint set has the advantages of equivalent optimization, good initialization, strong convergence, greedy monotonicity, and modular optimization. Based on this, the target complementary constraint optimization problem can be solved by using a smooth approximation method based on the minimization entropy function to obtain the initial health impact weight value and the differentiated health impact weight value, and obtain the target prediction model.

[0097] In this embodiment, after obtaining the target prediction model, the following steps are further included:

[0098] S400, acquiring target data, and predicting the target data based on the target prediction model to obtain a hypertension risk prediction result for a target patient, wherein the target data is hypertension-related data for the target patient.

[0099] After the target prediction model is trained through the aforementioned process, the target prediction model can be used to predict the risk of hypertension.

[0100] Similarly, after obtaining the blood pressure related data of the target patient, the target data is first preprocessed, and the discrete value attributes in the target data are converted into continuous value attributes to obtain the first target data. Then, the diastolic pressure and systolic pressure data in the first data are extracted, and after being divided into multiple target diastolic pressure intervals and multiple target systolic pressure intervals in a nonlinear manner, they are input into the target prediction model, and the target data is predicted based on the target interval impact weight value and the target differentiated health impact weight value to obtain the hypertension risk prediction result of the target patient. Specifically, in this embodiment, the hypertension risk prediction results of the target patient are divided into "low risk", "medium risk" and "high risk" levels, indicating the risk level of hypertension of the target patient.

[0101] Furthermore, after predicting the target data based on the target interval impact weight value and the target differentiated health impact weight value to obtain the hypertension risk prediction result of the target patient, it also includes:

[0102] generating a risk assessment report and a health management strategy based on the hypertension risk level of the target patient;

[0103] Dynamically monitor the real-time physiological data of the target patient and predict the hypertension risk level, so as to perform personalized health management for the target patient based on the latest hypertension risk level.

[0104] Specifically, after obtaining the hypertension risk level of the target patient, a risk report may also be generated, including:

[0105] The prediction results are integrated into an easy-to-understand report format, including risk levels and explanations of key influencing factors, to help doctors quickly understand the patient's high hypertension risk status and possible medical interventions.

[0106] Decision support can also be provided, including:

[0107] The quantitative analysis results provided can serve as the basis for formulating treatment plans, lifestyle adjustment recommendations and long-term health management strategies.

[0108] Furthermore, the target patient may be dynamically monitored and adjusted, including:

[0109] According to the regular examination results and model feedback of the target patients, the treatment plan is adjusted to achieve personalized and precise hypertension management.

[0110] This can assist healthcare professionals in making better decisions in hypertension risk assessment and management by providing an efficient and accurate tool.

[0111] In summary, this embodiment provides a hypertension risk assessment method based on historical data. When performing hypertension risk assessment on a target patient, an initial model containing multiple hidden layers is constructed, and then an initial data set including multiple hypertensive patients is obtained, the initial data set is preprocessed to obtain a target training set and a target support set, the initial model is trained based on the target training set, cross-validation and hyperparameter adjustment are performed based on the target support set, and a target prediction model is obtained. Finally, the target data is obtained, and the target data is predicted based on the target prediction model to obtain a hypertension risk prediction result for the target patient, wherein the target data is hypertension-related data for the target patient. The hypertension risk assessment method based on historical data provided in this embodiment uses a personal health score square mean evaluation algorithm with big data, nonlinearity, and medium-granularity blood pressure interval allocation, integrates multiple indicator features, constructs a target prediction model, and predicts the patient's hypertension risk based on a multi-level network model, so that the prediction accuracy is higher and the prediction results are more accurate.

[0112] It should be understood that, although the various steps in the flowcharts given in the accompanying drawings of the present invention are displayed in sequence according to the indications of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear description in this article, the execution of the steps in the present invention is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a part of the steps of the present invention may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these sub-steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.

[0113] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided by the present invention can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0114] Embodiment 2

[0115] Based on the above embodiments, the present invention also provides a hypertension risk assessment device based on historical data, and its functional module diagram is shown as follows: Figure 2 As shown, the hypertension risk assessment device based on historical data includes:

[0116] A model building module, used to build an initial model, wherein the initial model contains multiple hidden layers, as described in the first embodiment;

[0117] A data set acquisition module, used to acquire an initial data set, preprocess the initial data set to obtain a target training set and a target support set, wherein the initial data set includes data sets of multiple hypertensive patients, as specifically described in Example 1;

[0118] A training module, used to train the initial model based on the target training set, perform cross-validation and hyperparameter adjustment based on the target support set, and obtain a target prediction model, as specifically described in Embodiment 1;

[0119] The prediction module is used to obtain target data, predict the target data based on the target prediction model, and obtain the hypertension risk prediction result of the target patient. The target data is hypertension-related data of the target patient, as specifically described in Example 1.

[0120] Embodiment 3

[0121] Based on the hypertension risk assessment method based on historical data described in the first embodiment, the present invention further provides a terminal, whose principle block diagram can be as follows: Figure 3 The terminal includes a memory 10 and a processor 20. The memory 10 stores a hypertension risk assessment program based on historical data. When the processor 10 executes the computer program, at least the following steps can be implemented:

[0122] Constructing an initial model, wherein the initial model includes multiple hidden layers;

[0123] Acquire an initial data set, and preprocess the initial data set to obtain a target training set and a target support set, wherein the initial data set includes data sets of multiple hypertensive patients;

[0124] The initial model is trained based on the target training set, and cross-validation and hyperparameter adjustment are performed based on the target support set to obtain a target prediction model;

[0125] Target data is acquired, and the target data is predicted based on the target prediction model to obtain a hypertension risk prediction result of the target patient, wherein the target data is hypertension-related data of the target patient.

[0126] Wherein, constructing the initial model includes:

[0127] Based on the first formula, a deep learning neural network with K hidden layers is constructed;

[0128] The first formula is:

[0129]

[0130] Where X = H (0) is the input matrix of the original data, H(k) ,k=1,2,L,K is the k-th layer depth representation of X, W (k) is the weight of the kth layer of the initial model.

[0131] Wherein, the weight of the kth layer of the initial model is obtained based on the second formula;

[0132] The second formula is:

[0133]

[0134] Where V (k) is a random value matrix.

[0135] Wherein, the preprocessing of the initial data set includes:

[0136] Converting discrete value attributes in the initial data set into continuous value attributes to obtain a first data set;

[0137] Randomly extract 80% of the data in the first data set as an initial training set, and the remaining 20% ​​of the data in the first data set as an initial support set;

[0138] The blood pressure data in the initial training set and the initial support set are obtained, and the initial training set and the initial support set are grouped based on the blood pressure data to obtain the target training set and the target support set, wherein the blood pressure data are diastolic pressure and systolic pressure data.

[0139] The step of grouping the initial training set and the initial support set based on the blood pressure data to obtain the target training set and the target support set includes:

[0140] The diastolic pressure and systolic pressure in the initial training set and the initial support set are extracted and divided into multiple target intervals in a nonlinear manner;

[0141] The initial training set and the initial support set are grouped based on the target interval to obtain the target training set and the target support set.

[0142] The step of training the initial model based on the target training set includes:

[0143] Constructing an initial interval impact weight value corresponding to each target interval;

[0144] Constructing initial differentiated health impact weight values ​​within each target interval, wherein the differentiated health impact weight values ​​include weight values ​​corresponding to age, gender, race, region, body mass index, family medical history, dietary preferences, exercise frequency, mental health status, and pregnancy status;

[0145] After the initial model is trained based on the target training set, cross-validation and hyperparameter adjustment are performed based on the target support set to obtain the target interval impact weight value and the target differentiated health impact weight value.

[0146] The step of training the initial model based on the target training set includes:

[0147] For non-convex logistic regression loss functions, construct an optimization problem with target complementarity constraints;

[0148] The target complementarity constraint optimization problem is:

[0149]

[0150] stg(x)≥0,h(x)≥0,<g(x),h(x)> ≥0;

[0151] The target complementarity constraint optimization problem is solved by a smooth approximation method based on minimizing the entropy function.

[0152] Embodiment 4

[0153] The present invention also provides a storage medium storing one or more programs, which can be executed by one or more processors to implement the steps of the hypertension risk assessment method based on historical data described in the above embodiment.

[0154] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for assessing hypertension risk based on historical data, characterized in that: include: Constructing an initial model, wherein the initial model includes multiple hidden layers; Acquire an initial data set, and preprocess the initial data set to obtain a target training set and a target support set, wherein the initial data set includes data sets of multiple hypertensive patients; The initial model is trained based on the target training set, and cross-validation and hyperparameter adjustment are performed based on the target support set to obtain a target prediction model; Target data is acquired, and the target data is predicted based on the target prediction model to obtain a hypertension risk prediction result of the target patient, wherein the target data is hypertension-related data of the target patient.

2. The hypertension risk assessment method based on historical data according to claim 1, characterized in that: The constructing of the initial model comprises: Based on the first formula, a deep learning neural network with K hidden layers is constructed; The first formula is: Where X = H (0) is the input matrix of the original data, H (k) ,k=1,2,L,K is the k-th layer depth representation of X, W (k) is the weight of the kth layer of the initial model.

3. The hypertension risk assessment method based on historical data according to claim 2, characterized in that: Obtaining the weight of the kth layer of the initial model based on the second formula; The second formula is: Where V (k) is a random value matrix.

4. The hypertension risk assessment method based on historical data according to claim 1, characterized in that: The preprocessing of the initial data set includes: Converting discrete value attributes in the initial data set into continuous value attributes to obtain a first data set; Randomly extract 80% of the data in the first data set as an initial training set, and the remaining 20% ​​of the data in the first data set as an initial support set; The blood pressure data in the initial training set and the initial support set are obtained, and the initial training set and the initial support set are grouped based on the blood pressure data to obtain the target training set and the target support set, wherein the blood pressure data are diastolic pressure and systolic pressure data.

5. The method for assessing hypertension risk based on historical data according to claim 4, characterized in that: The grouping the initial training set and the initial support set based on the blood pressure data to obtain the target training set and the target support set includes: The diastolic pressure and systolic pressure in the initial training set and the initial support set are extracted and divided into multiple target intervals in a nonlinear manner; The initial training set and the initial support set are grouped based on the target interval to obtain the target training set and the target support set.

6. The method for assessing hypertension risk based on historical data according to claim 5, characterized in that: The training of the initial model based on the target training set includes: Constructing an initial interval impact weight value corresponding to each target interval; Constructing initial differentiated health impact weight values ​​within each target interval, wherein the differentiated health impact weight values ​​include weight values ​​corresponding to age, gender, race, region, body mass index, family medical history, dietary preferences, exercise frequency, mental health status, and pregnancy status; After the initial model is trained based on the target training set, cross-validation and hyperparameter adjustment are performed based on the target support set to obtain the target interval impact weight value and the target differentiated health impact weight value.

7. The hypertension risk assessment method based on historical data according to claim 1, characterized in that: The training of the initial model based on the target training set includes: For non-convex logistic regression loss functions, construct an optimization problem with target complementarity constraints; The target complementarity constraint optimization problem is: stg(x)≥0,h(x)≥0,<g(x),h(x)> ≥0; The target complementarity constraint optimization problem is solved by a smooth approximation method based on minimizing the entropy function.

8. A hypertension risk assessment device based on historical data, characterized in that: The device comprises: A model building module, used to build an initial model, wherein the initial model contains multiple hidden layers; A data set acquisition module is used to acquire an initial data set, preprocess the initial data set, and obtain a target training set and a target support set, wherein the initial data set includes data sets of multiple hypertensive patients; A training module, used to train the initial model based on the target training set, perform cross-validation and hyperparameter adjustment based on the target support set, and obtain a target prediction model; The prediction module is used to obtain target data, predict the target data based on the target prediction model, and obtain the hypertension risk prediction result of the target patient, wherein the target data is hypertension-related data of the target patient.

9. A terminal, characterized in that: The terminal includes: a processor, a storage medium communicatively connected to the processor, the storage medium being suitable for storing a plurality of instructions, and the processor being suitable for calling the instructions in the storage medium to execute the steps of the hypertension risk assessment method based on historical data as described in any one of claims 1-7.

10. A storage medium, characterized in that: The storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps of the hypertension risk assessment method based on historical data as described in any one of claims 1-7.