Prediction method, system and equipment for early stage of gestational diabetes mellitus and medium

Abstract

The invention belongs to the technical field of medical diagnosis and treatment, and discloses a pre-gestational diabetes prediction method, system, device and medium, and the method comprises the steps: obtaining the monitoring index data of a gestational population, the monitoring index data comprising the physiological information, living habits and environmental factors of the gestational population; constructing a gestational diabetes prediction model based on a radial basis function neural network model, and performing training optimization on the gestational diabetes prediction model based on the monitoring index data to obtain a trained gestational diabetes prediction model; and according to the trained gestational diabetes prediction model, executing gestational diabetes early-stage prediction of the to-be-predicted pregnant person, and outputting blood glucose level prediction data corresponding to the to-be-predicted pregnant person. According to the technical scheme, pre-disease prevention, early warning and prediction can be carried out on the gestational diabetes mellitus, and the pre-disease risk of the gestational diabetes mellitus can be reliably evaluated.

Description

Technical Field

[0001] The present invention belongs to the field of medical diagnosis and treatment technology, and in particular relates to a method, system, device and medium for predicting the early stage of gestational diabetes. Background Art

[0002] Gestational diabetes mellitus (GDM) refers to impaired glucose tolerance that occurs or is first discovered during pregnancy. It is a common pregnancy complication that significantly increases the risk of short-term or long-term metabolic diseases in pregnant women and their offspring. GDM has become an important public health issue. GDM not only increases the risk of complications in pregnant women, such as gestational hypertension, eclampsia, and ketoacidosis, but may also lead to neonatal complications such as hypoglycemia, hypocalcemia, and respiratory distress syndrome. Gestational diabetes mellitus (GDM) is a common condition that affects about 15% of pregnant women. High blood sugar in the mother can harm the baby and lead to problems such as premature birth and fetal arrest. Most women are diagnosed with GDM in the second trimester, but there is currently a lack of accurate prediction methods for early gestational diabetes at home and abroad. However, the pathogenesis of GDM has not yet been fully clarified, and the cause is not yet fully clarified. Traditional blood glucose monitoring and glucose tolerance tests have limitations in predicting the risk and outcome of GDM. Summary of the invention

[0003] The purpose of the present invention is to provide a method, system, device and medium for predicting the early stage of gestational diabetes to solve the problems existing in the above-mentioned prior art.

[0004] To achieve the above object, the present invention provides a method for predicting the early stage of gestational diabetes, comprising:

[0005] Acquiring monitoring indicator data of a pregnant population, wherein the monitoring indicator data includes physiological information, living habits and environmental factors of the pregnant population;

[0006] Constructing a gestational diabetes prediction model based on a radial basis neural network model, and training and optimizing the gestational diabetes prediction model based on the monitoring index data to obtain a trained gestational diabetes prediction model;

[0007] The pre-gestational diabetes prediction of the pregnant person to be predicted is performed according to the trained gestational diabetes prediction model, and the blood sugar level prediction data corresponding to the pregnant person to be predicted is output.

[0008] Optionally, the physiological information includes physiological indicators and family genetic history of the pregnant population.

[0009] Optionally, the living habits include eating habits and lifestyles of pregnant people.

[0010] Optionally, the training and optimization of the gestational diabetes prediction model based on the monitoring index data also includes data cleaning and standardization of the monitoring index data to obtain pre-processed monitoring index data, and dividing the pre-processed monitoring index data into an exposure group and a control group according to whether gestational diabetes occurs; calculating the disease risk ratio of each indicator in the corresponding exposure group and the corresponding control group; and training and optimizing the gestational diabetes prediction model based on each indicator and the corresponding disease risk ratio.

[0011] Optionally, the training and optimizing the gestational diabetes prediction model based on the monitoring index data specifically includes:

[0012] Designing an initial radial basis network structure, determining the number of input layer nodes and hidden layer neurons of the initial radial basis network structure, and determining the action function of the hidden unit;

[0013] After the initial radial basis network structure design is completed, the parameters are initialized, the central vector of the hidden unit is initialized using the clustering algorithm, and the weight of the output unit is initialized according to the disease risk ratio corresponding to each indicator;

[0014] The monitoring index data is input into the initial radial basis network structure after parameter initialization for classification prediction, and the training is carried out with the goal of minimizing the loss between the initial training results after classification prediction and the blood glucose level values ​​corresponding to each index to obtain a trained gestational diabetes prediction model.

[0015] A system for predicting early stages of gestational diabetes, comprising:

[0016] A data collection module is used to obtain monitoring index data of pregnant people, wherein the monitoring index data includes physiological information, living habits and environmental factors of pregnant people;

[0017] A model training module, used for constructing a gestational diabetes prediction model according to a radial basis neural network model, and training and optimizing the gestational diabetes prediction model based on the monitoring index data to obtain a trained gestational diabetes prediction model;

[0018] The risk prediction module is used to perform the pre-gestational diabetes prediction of the pregnant person to be predicted based on the trained gestational diabetes prediction model, and output the blood sugar level prediction data corresponding to the pregnant person to be predicted.

[0019] An electronic device comprises a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to execute a method for predicting an early stage of gestational diabetes.

[0020] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for predicting the early stage of gestational diabetes is implemented.

[0021] The technical effects of the present invention are:

[0022] The radial basis network used in the present invention can handle more complex patterns and relationships and provide better fitting capabilities. The present invention can prevent, warn and predict gestational diabetes before the disease occurs, reduce the probability of gestational diabetes, and has important significance and value for improving the health level of all mankind and benefiting the whole society. The prediction network designed by the present invention does not need to be trained with a large number of samples, and the network parameters are obtained according to theoretical analysis.

[0023] The present invention closely combines medicine and engineering, and utilizes computer technology, intelligent technology, and information technology to deeply integrate with the theoretical and clinical experience of medical personnel, forming a new network connecting the information data of pregnant personnel, thereby realizing full monitoring of the life process of this group of people, full management of their health status, and reliable assessment of the pre-disease risk of gestational diabetes. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0025] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0026] Figure 1 Schematic diagram of the RBF neural network structure of the gestational diabetes prediction model in an embodiment of the present invention;

[0027] Figure 2 This is a prediction flow chart in an embodiment of the present invention. DETAILED DESCRIPTION

[0028] Various exemplary embodiments of the present invention will now be described in detail. This detailed description should not be considered as limiting the present invention, but should be understood as a more detailed description of certain aspects, features, and embodiments of the present invention.

[0029] It should be understood that the terms described in the present invention are only for describing special embodiments and are not intended to limit the present invention. In addition, for the numerical range in the present invention, it should be understood that each intermediate value between the upper and lower limits of the scope is also specifically disclosed. Each smaller range between the intermediate value in any stated value or stated range and any other stated value or intermediate value in the described range is also included in the present invention. The upper and lower limits of these smaller ranges can be independently included or excluded in the scope.

[0030] Unless otherwise indicated, all technical and scientific terms used herein have the same meanings as those of ordinary skill in the art generally understood by the present invention. Although the present invention describes only preferred methods, any method similar or equivalent to that described herein may also be used in the implementation or testing of the present invention. All documents mentioned in this specification are incorporated by reference to disclose and describe the methods related to the documents. In the event of a conflict with any incorporated document, the content of this specification shall prevail.

[0031] It will be apparent to those skilled in the art that various modifications and variations may be made to the specific embodiments of the present invention description without departing from the scope or spirit of the present invention. Other embodiments derived from the present invention description will be apparent to those skilled in the art. The present application description and examples are exemplary only.

[0032] The words “include,” “including,” “have,” “contain,” etc. used in this article are open-ended terms, meaning including but not limited to.

[0033] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0034] Embodiment 1

[0035] like Figure 1 - Figure 2 As shown, this embodiment provides a method for predicting early stage of gestational diabetes, comprising:

[0036] Acquiring monitoring indicator data of a pregnant population, wherein the monitoring indicator data includes physiological information, living habits and environmental factors of the pregnant population;

[0037] Constructing a gestational diabetes prediction model based on a radial basis neural network model, and training and optimizing the gestational diabetes prediction model based on the monitoring index data to obtain a trained gestational diabetes prediction model;

[0038] The pre-gestational diabetes prediction of the pregnant person to be predicted is performed according to the trained gestational diabetes prediction model, and the blood sugar level prediction data corresponding to the pregnant person to be predicted is output.

[0039] The pre-gestational diabetes prediction model realizes the comprehensive digitization and networking of the population's comprehensive information through various monitoring, data entry, analysis and other sensing means. It closely combines computer technology, sensing technology, transmission technology, intelligent technology, information technology, modern information management technology, etc. with the conditions of pregnant people, forming a new network connecting the information data of pregnant people, realizing the full monitoring of the life process of this population, the full management of health status and the reliable assessment of the pre-disease risk of gestational diabetes.

[0040] (1) Research on the mechanism and pathology of gestational diabetes

[0041] The cause and onset of gestational diabetes is a complex process affected by multiple factors, which requires detailed analysis and research. Since there are many factors affecting gestational diabetes and many characteristic quantities reflecting its pathological signs, it is necessary to explore the pathogenesis of gestational diabetes, study its pathological mechanism, and deeply explore the relationship and influence between diabetes and other diseases, and find quantitative indicators that affect the pathogenesis of gestational diabetes and various influencing factors.

[0042] There are many factors that cause gestational diabetes, and the degree of impact varies greatly. Some are direct, some are indirect, some are serious, some are general, some are far-reaching, and some are short-term. Therefore, when conducting a mechanism and pathological analysis of gestational diabetes, it is necessary not only to find the influencing factors of gestational diabetes, but also to distinguish between the primary and secondary causes, and to break down the specific numbers of the impact weights to determine the importance and severity of these indicators in the development of gestational diabetes.

[0043] Based on the above analysis, multiple physical indicators of pregnant women are monitored so that by analyzing and integrating these monitoring indicators, predictions can be made on whether and when pregnant women will develop diabetes, and precautions can be given to avoid its occurrence as much as possible.

[0044] (2) Predictive monitoring indicators for gestational diabetes

[0045] Through in-depth analysis of the pathological mechanism of gestational diabetes, we found various factors that determine and influence the occurrence and development of gestational diabetes, such as age, weight, BMI (weight / height) and other factors related to one's own conditions. 2 kg / m 2 ), waist circumference, hip circumference, fasting blood sugar, blood sugar two hours after meal, HbA1c, blood pressure, blood lipids, serum; related to family inheritance: paternal ancestors, father, maternal ancestors, mother; and lifestyle habits: like sweets, heavy taste, dinner time, smoking, ...; etc., as listed in Table 1, where the OR value is how many times the risk of disease in the exposure group is the risk of disease in the control group. The monitoring index is used as the input of the neural network. The input vector of this embodiment is X(x1, x2, ..., xp ), where p = 103, when the normal value is 0, the abnormal value is 1. When the normal value is 0, the abnormal value is 1.

[0046] Table 1 List of parameters of each indicator in the prediction model of gestational diabetes

[0047]

[0048]

[0049]

[0050]

[0051]

[0052] (3) Design of prediction model for gestational diabetes

[0053] The radial basis function (RBF) network is a regularized network with only one hidden layer. The action function of the hidden unit is the Green function, and the weights from the hidden layer to the output unit are determined by G and λ. From the perspective of function approximation, the regularized network is a universal approximator. As long as there are enough hidden units, it can arbitrarily approximate the function defined in R. p Any multivariate continuous function on a compact set of ; since the output is linear with respect to the unknown coefficients, it has the best approximation performance, that is, given an unknown nonlinear function f, a set of coefficients can always be selected that makes the approximation to f better than other possible choices; the solution obtained by the regularized network is optimal, that is, it minimizes the oscillation of the function, so that the result will not be consistent with the data at a given sample point, but will oscillate at other points.

[0054] Since the radial basis function (RBF) network is a regularized network, it has the above advantages. As long as there are enough hidden neurons, any single-valued continuous function can be approximated with arbitrary accuracy regardless of the form of the radial basis function. The network has been theoretically proven to provide the approximation capability of a general three-layer network.

[0055] Use the Gaussian function:

[0056]

[0057] Gaussian function is used as the basis function of radial basis neural network, where p is the dimension of input vector, that is, the number of inputs, thus:

[0058]

[0059] Where N is the number of hidden layer neurons. Thus, the parameters of the radial basis neural network include the output unit weight ω i , the center x of the hidden uniti and basis function width σ .

[0060] RBF neural network prediction model for gestational diabetes mellitus Figure 1 As shown. Through analysis and summary, the number of input samples is p, in this embodiment p = 103; the number of input-input samples is N, the number of hidden units is N, then the center vector of the hidden unit is X i , basis function width σ The number of output unit weights ω is also N. In this embodiment, N=50, which is the same as the type of samples. The input sample pairs are the monitoring values ​​of the indicators in Table 1, and the output samples are the blood sugar level values. The center X of the hidden unit is obtained by clustering method. i , using the OR value weight as the weight ω i Finally, the least square method is used to σ After training, f(x) is the predicted result, that is, the risk probability of gestational diabetes.

[0061] The simulation platform for implementing this embodiment is Matlab, which has a radial basis neural network model. As long as the number of inputs, the function and number of hidden layers are set as required, the input-output sample data is input, and training is performed according to the least squares method.

[0062] This embodiment uses a radial basis network, which can handle more complex patterns and relationships and provide better fitting capabilities. Compared with other neural networks, the prediction network designed in this embodiment does not require a large number of sample training, and the network parameters are obtained based on theoretical analysis.

[0063] This embodiment can prevent, warn and predict gestational diabetes before it occurs, and reduce the probability of gestational diabetes, which is of great significance and value to improving the health level of all mankind and benefiting the whole society.

[0064] This embodiment closely combines medicine and engineering, and uses computer technology, intelligent technology, and information technology to deeply integrate with the theoretical and clinical experience of medical personnel, forming a new network that connects the information data of pregnant people, realizing full monitoring of the life process of this group of people, full management of their health status, and reliable assessment of the pre-disease risk of gestational diabetes.

[0065] A system for predicting early stages of gestational diabetes, comprising:

[0066] A data collection module is used to obtain monitoring index data of pregnant people, wherein the monitoring index data includes physiological information, living habits and environmental factors of pregnant people;

[0067] A model training module, used for constructing a gestational diabetes prediction model according to a radial basis neural network model, and training and optimizing the gestational diabetes prediction model based on the monitoring index data to obtain a trained gestational diabetes prediction model;

[0068] The risk prediction module is used to perform the pre-gestational diabetes prediction of the pregnant person to be predicted based on the trained gestational diabetes prediction model, and output the blood sugar level prediction data corresponding to the pregnant person to be predicted.

[0069] An electronic device comprises a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to execute a method for predicting an early stage of gestational diabetes.

[0070] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for predicting the early stage of gestational diabetes is implemented.

[0071] The above is only a preferred specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for predicting early stage of gestational diabetes, characterized in that: include: Acquiring monitoring indicator data of a pregnant population, wherein the monitoring indicator data includes physiological information, living habits and environmental factors of the pregnant population; Constructing a gestational diabetes prediction model based on a radial basis neural network model, and training and optimizing the gestational diabetes prediction model based on the monitoring index data to obtain a trained gestational diabetes prediction model; The pre-gestational diabetes prediction of the pregnant person to be predicted is performed according to the trained gestational diabetes prediction model, and the blood sugar level prediction data corresponding to the pregnant person to be predicted is output.

2. A method for predicting early stage of gestational diabetes according to claim 1, characterized in that: The physiological information includes physiological indicators and family genetic history of the pregnant population.

3. A method for predicting early stage of gestational diabetes according to claim 1, characterized in that: The living habits include the eating habits and lifestyle of pregnant people.

4. The method for predicting early stage of gestational diabetes according to claim 1, characterized in that: The method further includes performing data cleaning and standardization on the monitoring index data to obtain pre-processed monitoring index data, and dividing the pre-processed monitoring index data into an exposure group and a control group according to whether gestational diabetes occurs; calculating the disease risk ratio of each index in the corresponding exposure group and the corresponding control group; and training and optimizing the gestational diabetes prediction model based on each index and the corresponding disease risk ratio.

5. The method for predicting early stage of gestational diabetes according to claim 1, characterized in that: The training and optimization of the gestational diabetes prediction model based on the monitoring index data specifically includes: Designing an initial radial basis network structure, determining the number of input layer nodes and hidden layer neurons of the initial radial basis network structure, and determining the action function of the hidden unit; After the initial radial basis network structure design is completed, the parameters are initialized, the central vector of the hidden unit is initialized using the clustering algorithm, and the weight of the output unit is initialized according to the disease risk ratio corresponding to each indicator; The monitoring index data is input into the initial radial basis network structure after parameter initialization for classification prediction, and the training is carried out with the goal of minimizing the loss between the initial training results after classification prediction and the blood glucose level values ​​corresponding to each index to obtain a trained gestational diabetes prediction model.

6. A system for predicting early stages of gestational diabetes, characterized in that: include: A data collection module is used to obtain monitoring index data of pregnant people, wherein the monitoring index data includes physiological information, living habits and environmental factors of pregnant people; A model training module, used for constructing a gestational diabetes prediction model according to a radial basis neural network model, and training and optimizing the gestational diabetes prediction model based on the monitoring index data to obtain a trained gestational diabetes prediction model; The risk prediction module is used to perform the pre-gestational diabetes prediction of the pregnant person to be predicted based on the trained gestational diabetes prediction model, and output the blood sugar level prediction data corresponding to the pregnant person to be predicted.

7. An electronic device, characterized in that: The electronic device comprises a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to perform a method for predicting an early stage of gestational diabetes according to any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that: The device stores a computer program, which, when executed by a processor, implements a method for predicting the early stage of gestational diabetes as described in any one of claims 1 to 5.

Images