Type 2 diabetes prediction method and system based on medical data processing

Abstract

The invention belongs to the technical field of diabetes prediction, and particularly relates to a type 2 diabetes prediction method and system based on medical data process.The method comprises the steps that when medical data is processed, especially a medical data sample set is processed, firstly, data complementation operation is conducted on samples with data missing, and then data complementation operation is conducted on the samples with data missing; different data complementing methods are adopted for different sample sets, complementing operation is carried out on samples with accurate comparable samples in a similarity comparison mode, complementing operation is carried out on samples without comparable samples in a deep learning mode, in this way, the accuracy of medical data processing is improved, and meanwhile the accuracy of medical data processing is improved. And meanwhile, after data complementation, a quality detection link is also set, that is, complemented samples are input into the type 2 diabetes mellitus prediction model trained by samples which are not complemented, and the data complementation effect is judged according to the output result of the model, so that the data complementation effect is more intuitive and easy to compete.

Description

Technical Field

[0001] The present invention belongs to the technical field of diabetes prediction, and more specifically, relates to a method and system for predicting type 2 diabetes based on medical data processing. Background Art

[0002] Type 2 diabetes mellitus (T2DM) is a chronic disease caused by insufficient insulin use or reduced efficiency. It is common in adults and is also called adult-onset diabetes. The disease is caused by genetics and environmental factors such as lifestyle, overnutrition, and lack of physical activity. The initial symptoms are often mild, and many people are not discovered until complications occur or during routine physical examinations. Since type 2 diabetes has many clinical complications, it is easy for patients to develop cardiovascular and cerebrovascular diseases. In severe cases, it can also cause symptoms in the limbs and multiple parts of the body. Therefore, the prediction of type 2 diabetes has very important practical significance.

[0003] Machine learning (ML) is a subfield of artificial intelligence. It can identify and extract features of labeled training set data, build prediction models, and use the learned rules to predict and classify new data. It has been widely used in the medical field. ML models can handle missing data and nonlinear relationships between variables more flexibly, which is more conducive to the diagnosis and prognosis of diseases. Clinicians can use this to make further and timely judgments on patients' conditions.

[0004] There are schemes for predicting type 2 diabetes based on machine learning in the prior art. For example, a Chinese invention patent (CN106384013A) discloses a method for predicting type 2 diabetes based on an incremental neural network model, which includes the following steps: establishing a daily data database for type 2 diabetes; training a neural network model; collecting daily life data and sending it to a server, and saving it in a user's daily data record table; extracting the data of the day from the user's daily data record table to form an n-dimensional vector, normalizing it, and inputting it into a type 2 diabetes pathology neural network model for diabetes probability prediction; a smart home diabetes care device determines whether the diabetes probability value is greater than 0.5; when the user is determined to have type 2 diabetes, the user goes to the hospital for examination by himself, and transmits the examination results back to the server through the smart home diabetes care device, and the server determines whether the examination results are correct; when the examination results are wrong, an incremental algorithm is executed to dynamically correct the neural network model. The above diagnostic results focus on establishing a prediction model for diabetes prediction, and do not pay enough attention to the data processing link, resulting in problems with data processing accuracy, which affects the accuracy of type 2 diabetes diagnosis; at the same time, the establishment process of the deep learning model in the prior art is relatively rough, resulting in low efficiency in solving problems by the model. Summary of the invention

[0005] The purpose of the present invention is to provide a method and system for predicting type 2 diabetes based on medical data processing, so as to improve the accuracy of type 2 diabetes diagnosis.

[0006] In order to solve the above technical problems, the present invention specifically provides the following technical solutions: a method for predicting type 2 diabetes based on medical data processing, specifically comprising the steps of: S1: Obtain a sample set for type 2 diabetes prediction; S2: performing a first data preprocessing operation on the sample set to obtain a second sample set; S3: performing a second data preprocessing operation on the second sample set to obtain a third sample set; The S3 is specifically as follows: S3.1: traverse the second sample set, and mark the samples in the second sample set that have missing index values; S3.2: judge the index of the missing value in the sample, if it is a basic index, do not perform the supplement operation, if it is a laboratory index, proceed to the next step; S3.3: supplement the laboratory index value; The S3.3 is specifically: S3.3.1: Determine the similarity between the sample to be supplemented and other samples in the second sample set. If there are other samples with a similarity of more than 80%, proceed to the next step, otherwise proceed to S3.3.3; S3.3.2: Calculate the value of the index to be supplemented in the sample to be supplemented according to the value of the index to be supplemented in other samples with a similarity of more than 80%; S3.3.3: Use the second supplementation method to perform data correction operation on the sample to be supplemented; S4: performing quality detection on the samples in the third sample set that have undergone the supplementation operation to obtain a fourth sample set; S4.1: Establish a type 2 diabetes prediction model; S4.2: Use the samples in the third sample set that have not been supplemented to train the type 2 diabetes prediction model; S4.3: Input the samples in the third sample set that have been supplemented to the type 2 diabetes prediction model; Output the probability of type 2 diabetes; S4.4: Judge the output probability of type 2 diabetes, if it is greater than a preset value, retain it, otherwise delete the sample, and obtain a fourth sample set; S5: retraining the type 2 diabetes prediction model with the fourth sample set, inputting the patient data to be predicted into the trained type 2 diabetes prediction model to obtain a prediction result.

[0007] Preferably, in S1, the samples in the sample set include: basic indicators of patients and laboratory indicators.

[0008] Preferably, basic indicators include the patient's age, gender, height, weight, whether there is polyuria, whether there is polydipsia, whether there is physical weakness, whether there is polyphagia, whether there is blurred vision, whether there is skin itching, whether there is irritability, and whether there is obesity; laboratory indicators include: fasting blood sugar, postprandial blood sugar, glycosylated hemoglobin, blood lipids, blood pressure, meal 2hC-peptide, blood urea nitrogen, and fasting C-peptide.

[0009] Preferably, in S2, the first data preprocessing operation is: sample screening and feature encoding.

[0010] Preferably, in S3.3.1, the similarity comparison is specifically as follows: first, obtain a sample with the same gender as the sample to be supplemented and an age difference of less than 5 years as the sample to be supplemented as a comparison sample, calculate the similarity between the laboratory indicator values ​​of the sample to be supplemented except for the indicator to be supplemented and the laboratory indicator values ​​of the same type in the comparison sample, and then average the multiple similarities, and realize the similarity comparison between the sample to be supplemented and other samples based on the above similarity average value.

[0011] Preferably, S3.3.2 specifically includes: calculating the average value of the index to be supplemented in other samples with a similarity exceeding 80% as the value of the index to be supplemented in the sample to be supplemented.

[0012] Preferably, in S3.3.3, the second supplementation method is to use a deep learning model to perform data correction operations on the samples to be supplemented.

[0013] Preferably, the deep learning model is a BP neural network model.

[0014] Preferably, the formula for determining the number M of neurons in the hidden layer of the BP neural network model is:

[0015] Where M s is the number of sample sets of the BP neural network model, M o is the number of neurons in the output layer of the BP neural network model, M i is the number of neurons in the input layer of the BP neural network model, and a is the adjustment coefficient, which is within the range of 3-5.

[0016] According to another aspect of the present invention, a type 2 diabetes prediction system based on medical data processing is provided, wherein the system adopts the above-mentioned type 2 diabetes prediction method based on medical data processing, and the system further comprises: A sample set acquisition module, which acquires a sample set for type 2 diabetes prediction; A first preprocessing module, configured to perform a first data preprocessing operation on the sample set to obtain a second sample set; A second preprocessing module, configured to perform a second data preprocessing operation on the second sample set to obtain a third sample set; A quality detection module, used for performing quality detection on the samples in the third sample set that have undergone the supplementation operation, to obtain a fourth sample set; The type 2 diabetes prediction model is used to retrain the type 2 diabetes prediction model with the fourth sample set, input the patient data to be predicted into the trained type 2 diabetes prediction model to obtain a prediction result.

[0017] Compared with the prior art, the present invention has the following beneficial effects: When processing medical data, especially when processing medical data sample sets, the present invention first performs data supplementation operations on samples with missing data, adopts different data supplementation methods for different sample sets, and adopts a similarity comparison method to perform supplementation operations on samples with accurate control samples, and adopts a deep learning method to perform supplementation operations on samples without control samples. In this way, the accuracy of medical data processing is improved, while taking into account the prediction efficiency. At the same time, after the data is supplemented, a quality inspection link is set up, that is, the supplemented samples are input into the type 2 diabetes prediction model trained with unsupplemented samples, and the data supplementation effect is judged according to the output results of the model, making the data supplementation effect more intuitive and easy to convince; At the same time, the present invention proposes a new parameter determination formula for the BP neural network model. Through the above formula, a more accurate number of hidden layer neurons can be obtained, thereby improving the efficiency of the BP neural network model in medical data supplementation. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the implementation methods of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for the implementation methods or the description of the prior art. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other implementation drawings can be derived from the provided drawings without creative work.

[0019] Figure 1 A flowchart of a method for predicting type 2 diabetes based on medical data processing provided by an embodiment of the present invention; Figure 2 A flowchart of performing a second data preprocessing operation on the second sample set to obtain a third sample set provided by an embodiment of the present invention; Figure 3 A flowchart of supplementing the laboratory index value provided by an embodiment of the present invention; Figure 4A flow chart of performing quality detection on the samples in the third sample set that have undergone a supplementation operation to obtain a fourth sample set is provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0020] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] The following first describes the concepts involved in the present application in conjunction with the accompanying drawings. It should be noted that the following description of each concept is only to make the content of the present application easier to understand, and does not limit the scope of protection of the present application; at the same time, the embodiments and features in the embodiments of the present application can be combined with each other in the absence of conflict. The present application will be described in detail with reference to the accompanying drawings and in conjunction with the embodiments.

[0022] like Figure 1 As shown, the present invention provides a method for predicting type 2 diabetes based on medical data processing, which specifically includes the steps of: S1: Obtain a sample set for type 2 diabetes prediction; In this step, 4459 patients diagnosed with type 2 diabetes (T2DM) were selected as sample subjects by filling out questionnaires, including 2912 males and 1547 females, aged 51-79 years, with an average age of 63.6±3.5 years, and a BMI index of 16-28 kg / m 2 , average BMI 25.13+2.13 kg / m 2 ; The sample collection process strictly abides by ethical principles to ensure patient privacy and rights; Among them, when obtaining a sample set for type 2 diabetes prediction, some irrelevant samples need to be excluded, such as patients with unclear diabetes types and special types of diabetes, and the sample objects are all diagnosed for the first time; In this step, the samples in the sample set include: basic indicators of patients and laboratory indicators; Among them, the basic indicators include the patient's age, gender, height, weight, whether he has polyuria, polydipsia, physical weakness, polyphagia, blurred vision, skin itching, irritability, obesity, etc. Laboratory indicators include: fasting blood glucose, postprandial blood glucose, glycosylated hemoglobin, blood lipids, blood pressure, meal 2hC-peptide, blood urea nitrogen, fasting C-peptide, etc.; S2: performing a first data preprocessing operation on the sample set to obtain a second sample set; Wherein, the first data preprocessing operation is: sample screening and feature coding; Real medical data generally has problems such as irregular data format, missing values ​​and many outliers. Therefore, it is necessary to first screen the samples. The sample screening is specifically as follows: if there are more than two unqualified indicators in the sample, the sample is identified as an unqualified sample, and then the unqualified sample is deleted, and all the samples are judged as unqualified, thereby completing the sample screening; Among them, the unqualified indicators are indicators with irregular data format, missing values, and abnormal values; Machine learning models usually require input to be in numerical form. Therefore, it is necessary to perform feature encoding on the features in the sample. In this step, the LabelEncoder method is used to map each non-numerical indicator to a numerical value. In the gender indicator, male is 1 and female is 0. In the indicators of whether there is polyuria, polydipsia, physical weakness, polyphagia, blurred vision, itchy skin, irritability, and obesity, yes is 1 and no is 0. S3: performing a second data preprocessing operation on the second sample set to obtain a third sample set; In the above steps, samples with more than two unqualified indicators were deleted. There are still some samples in the sample set with missing indicator values. Therefore, the sample set continues to be preprocessed to improve the accuracy of data processing; In this step, if Figure 2 As shown, the S3 is specifically: S3.1: traverse the second sample set, and mark the samples in the second sample set that have missing index values; In this step, in the second sample set, if there is a missing value of an indicator, it will generally be displayed in the form of "NULL" in the sample set. Therefore, in this step, by searching the "NULL" keyword in the sample set, the indicator with missing value can be obtained; S3.2: Determine the missing value indicator in the sample. If it is a basic indicator, no supplementation operation is performed. If it is a laboratory indicator, proceed to the next step. Among them, for basic indicators, age, gender, height, weight, polyuria, polydipsia, physical weakness, polyphagia, blurred vision, skin itching, irritability, obesity, these cannot be supplemented, so no supplementation operation is performed in this step; for laboratory indicators, they are all numerical indicators and can be supplemented; S3.3: Supplement the laboratory index values; For medical data, there is no reference between the data of the same indicator of different samples. Therefore, the operation of taking the average or mode of adjacent values ​​in the prior art will cause a large supplementation error. Therefore, this embodiment proposes a method for supplementing missing data in a medical data sample set. As attached Figure 3 As shown, the S3.3 is specifically: S3.3.1: Determine the similarity between the sample to be supplemented and other samples in the second sample set. If there are other samples with a similarity exceeding 80%, proceed to the next step; otherwise proceed to S3.3.3; As a preferred embodiment, the similarity comparison is specifically as follows: firstly, a sample of the same gender as the sample to be supplemented and with an age difference of less than 5 years is obtained as a comparison sample, and the similarity between the laboratory index value of the sample to be supplemented except the index to be supplemented and the laboratory index value of the same type in the comparison sample is calculated, and then the average value of the multiple similarities is calculated, and the similarity comparison between the sample to be supplemented and other samples is realized based on the above similarity average value; S3.3.2: Calculate the value of the indicator to be supplemented in the sample to be supplemented according to the value of the indicator to be supplemented in other samples with a similarity exceeding 80%; In this step, S3.3.2 is specifically: calculating the average value of the index to be supplemented in other samples with a similarity exceeding 80% as the value of the index to be supplemented in the sample to be supplemented; In this step, through the above operations, firstly, patient samples of the same gender and similar age are screened out, and then samples with similar physical conditions as the patients of the samples to be supplemented are screened out through similarity comparison, and then the missing index values ​​in the samples to be supplemented are supplemented according to the physical index values ​​of the samples with similar physical conditions, that is, a relatively accurate supplementation effect can be achieved by using a relatively simple method; S3.3.3: Performing data correction operation on the sample to be supplemented using the second supplementation method; Wherein, the second supplementation method is to use a deep learning model to perform a data correction operation on the sample to be supplemented; Wherein, the deep learning model is a BP neural network model; It is worth noting that the input of the deep learning model is the sample to be supplemented, and the output of the deep learning is the value of the indicator to be supplemented; In the steps, the BP neural network model includes models for supplementing different indicators, that is, if the fasting blood glucose value is missing in the sample to be supplemented, the BP neural network model for supplementing the fasting blood glucose value is adopted; if the postprandial blood glucose value is missing in the sample to be supplemented, the BP neural network model for supplementing the fasting blood glucose value is adopted; if the glycosylated hemoglobin value is missing in the sample to be supplemented, the BP neural network model for supplementing the fasting blood glucose value is adopted, etc.; If the fasting blood glucose value and the postprandial blood glucose value are missing in the sample to be supplemented, the BP neural network model for supplementing the fasting blood glucose value and the postprandial blood glucose value is used; if the fasting blood glucose value and the blood lipid value are missing in the sample to be supplemented, the BP neural network model for supplementing the fasting blood glucose value and the blood lipid value is used, etc.; Among them, different types of BP neural networks have the same structure, and the difference is that different training sets are used to train the BP neural network model. For example, the BP neural network model that supplements the fasting blood glucose value uses the training set of the missing fasting blood glucose value to train the BP neural network model; At the same time, when the BP neural network model is established, whether it is a regression or classification task, the selection of an appropriate number of hidden layer neurons will greatly affect the performance of the BP neural network. In theory, the more hidden layer neurons there are, the more accurate the calculation results will be. In fact, for simple data sets, a single-digit number of hidden layer neurons can generally meet the calculation requirements, but for complex data sets, a larger number of hidden layer neurons is required. However, the more hidden layer neurons there are, the more complicated the calculation process will become, requiring more powerful hardware equipment to support the operation of the model, and the operation time will also increase in an exponential form. Generally, the number of neurons in the hidden layer is determined by the number of neurons in the input layer and the number of neurons in the output layer. In view of the characteristics of medical data and the relatively small number of sample sets, this embodiment determines the number of hidden layer neurons in the BP neural network based on the following formula; The formula for determining the number of neurons M in the hidden layer of the BP neural network model is:

[0023] Where M s is the number of sample sets of the BP neural network model, M o is the number of neurons in the output layer of the BP neural network model, M i is the number of neurons in the input layer of the BP neural network model, and a is the adjustment coefficient, which is generally within the range of 3-5; Through the above formula, a more accurate number of hidden layer neurons can be obtained, thereby improving the efficiency of the BP neural network model in medical data supplementation; compared with the model mentioned in the background technology, the above BP neural network model is used to train and solve the model on the same server when the data set and other hyperparameters of the model are the same. The model training and solution time provided by this embodiment are shorter than the incremental neural network model mentioned in the background technology, showing a greater model performance.

[0024] S4: performing quality detection on the samples in the third sample set that have undergone the supplementation operation to obtain a fourth sample set; Through the above steps, the samples to be supplemented can be supplemented. For medical data, if there are inaccurate samples, it will have a greater impact on the subsequent prediction of type 2 diabetes. Therefore, this embodiment sets a step of quality testing on the samples that have been supplemented. Among them, as attached Figure 4 As shown, the S4 is specifically: S4.1: Establish a prediction model for type 2 diabetes; Wherein, the type 2 diabetes prediction model is a deep learning model, specifically a BP neural network model; S4.2: training the type 2 diabetes prediction model using the samples in the third sample set that have not been supplemented; In this step, all samples in the third sample set that are not used for the supplementation operation are used to train the type 2 diabetes prediction model, and the next step is entered after the training is completed; S4.3: inputting the samples in the third sample set that have undergone the supplementation operation into the type 2 diabetes prediction model; and outputting the probability of type 2 diabetes; Among them, after the training in the above steps, the type 2 diabetes prediction model already has prediction capability, therefore, this embodiment uses the type 2 diabetes prediction model to detect the complementary quality of the sample set; S4.4: judging the output probability of type 2 diabetes, if it is greater than a preset value, retaining it, otherwise deleting the sample, and obtaining a fourth sample set; Among them, in the quality inspection link, this embodiment inputs the supplemented samples into the type 2 diabetes prediction model in the training process for quality inspection. If the probability of disease output by the prediction model is greater than the preset value, it means that the data in the sample is supplemented more accurately. Therefore, the sample is retained. If the probability of disease output by the prediction model is less than the preset value, it means that the data in the sample is not supplemented accurately, resulting in an erroneous conclusion output in the sample input prediction model. Therefore, the sample needs to be deleted to improve the accuracy of the sample set.

[0025] In this embodiment, in order to improve the accuracy of the sample set, the preset value is 70%.

[0026] When processing medical data, especially when processing medical data sample sets, the present invention first performs data supplementation operations on samples with missing data, adopts different data supplementation methods for different sample sets, and adopts a similarity comparison method to perform supplementation operations on samples with accurate comparable samples, and adopts a deep learning method to perform supplementation operations on samples without comparable samples, thereby improving the accuracy and efficiency of medical data processing; At the same time, after the data is supplemented, a quality inspection link is set up, that is, the supplemented samples are input into the type 2 diabetes prediction model trained with unsupplemented samples, and the data supplementation effect is judged according to the output results of the model, making the data supplementation effect more intuitive and convincing.

[0027] S5: retraining the type 2 diabetes prediction model with the fourth sample set, inputting the patient data to be predicted into the trained type 2 diabetes prediction model to obtain a prediction result.

[0028] The prediction result is the probability of the patient to be predicted suffering from type 2 diabetes.

[0029] Embodiment 2: This embodiment further includes a type 2 diabetes prediction system based on medical data processing, the system adopts a type 2 diabetes prediction method based on medical data processing of embodiment 1, and the system further includes: A sample set acquisition module, which acquires a sample set for type 2 diabetes prediction; A first preprocessing module, configured to perform a first data preprocessing operation on the sample set to obtain a second sample set; A second preprocessing module, configured to perform a second data preprocessing operation on the second sample set to obtain a third sample set; A quality detection module, used for performing quality detection on the samples in the third sample set that have undergone the supplementation operation, to obtain a fourth sample set; The type 2 diabetes prediction model is used to retrain the type 2 diabetes prediction model with the fourth sample set, input the patient data to be predicted into the trained type 2 diabetes prediction model to obtain a prediction result.

[0030] Embodiment three, this embodiment includes an electronic device, including a processor and a memory, wherein the memory stores a program or instruction that can be run on the processor, and when the program or instruction is executed by the processor, each step of a method for predicting type 2 diabetes based on medical data processing in embodiment one is implemented, and the same technical effect can be achieved.

[0031] Embodiment 4: This embodiment includes a computer-readable storage medium on which a data processing program is stored. The data processing program is executed by a processor to perform various steps of a method for predicting type 2 diabetes based on medical data processing in embodiment 1.

[0032] It should be understood by those skilled in the art that the embodiments of this article can be provided as methods, devices (equipment), or computer program products. Therefore, this article can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Including but not limited to RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disk (DVD) or other optical disk storage, magnetic cassette, tape, disk storage or other magnetic storage device, or any other medium that can be used to store desired information and can be accessed by a computer. In addition, it is well known to those of ordinary skill in the art that communication media generally contain computer-readable instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0033] This document is described with reference to flowcharts and / or block diagrams of methods, apparatuses (devices), and computer program products according to the embodiments of this document. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0034] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0035] The embodiments and / or implementation methods described above are only used to illustrate the preferred embodiments and / or implementation methods of the technology of the present invention, and are not intended to impose any formal restrictions on the implementation methods of the technology of the present invention. Any person skilled in the art may make some changes or modifications to other equivalent embodiments without departing from the scope of the technical means disclosed in the content of the present invention, but they should still be regarded as the technology or embodiments that are essentially the same as the present invention. Note that the above are only preferred embodiments of the present invention and the technical principles used. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described here, and that various obvious changes, readjustments and substitutions can be made to those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention is described in more detail through the above embodiments, the present invention is not limited to the above embodiments, and may also include more other equivalent embodiments without departing from the concept of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. A method for predicting type 2 diabetes based on medical data processing, characterized in that: The specific steps include: S1: Obtain a sample set for type 2 diabetes prediction; S2: performing a first data preprocessing operation on the sample set to obtain a second sample set; S3: performing a second data preprocessing operation on the second sample set to obtain a third sample set; The S3 is specifically as follows: S3.1: traverse the second sample set, and mark the samples in the second sample set that have missing index values; S3.2: judge the index of the missing value in the sample, if it is a basic index, do not perform the supplement operation, if it is a laboratory index, proceed to the next step; S3.3: supplement the laboratory index value; The S3.3 is specifically as follows: S3.3.1: determine the similarity between the sample to be supplemented and other samples in the second sample set, if there are other samples with a similarity exceeding 80%, proceed to S3.3.2, otherwise proceed to S3.3.3; S3.3.2: calculate the value of the index to be supplemented in the sample to be supplemented according to the value of the index to be supplemented in other samples with a similarity exceeding 80%; S3.3.3: use the second supplementation method to perform data correction operation on the sample to be supplemented; S4: performing quality detection on the samples in the third sample set that have undergone the supplementation operation to obtain a fourth sample set; S4.1: Establish a type 2 diabetes prediction model; S4.2: Use the samples in the third sample set that have not been supplemented to train the type 2 diabetes prediction model; S4.3: Input the samples in the third sample set that have been supplemented to the type 2 diabetes prediction model; Output the probability of type 2 diabetes; S4.4: Judge the output probability of type 2 diabetes, if it is greater than a preset value, retain it, otherwise delete the sample, and obtain a fourth sample set; S5: retraining the type 2 diabetes prediction model using the fourth sample set, inputting the patient data to be predicted into the trained type 2 diabetes prediction model to obtain a prediction result.

2. The method for predicting type 2 diabetes based on medical data processing according to claim 1, characterized in that: In S1, the samples in the sample set include: basic indicators of patients and laboratory indicators.

3. The method for predicting type 2 diabetes based on medical data processing according to claim 2, characterized in that: Basic indicators include the patient's age, gender, height, weight, whether there is polyuria, polydipsia, physical weakness, polyphagia, blurred vision, skin itching, irritability and obesity; laboratory indicators include: fasting blood sugar, postprandial blood sugar, glycosylated hemoglobin, blood lipids, blood pressure, meal 2hC-peptide, blood urea nitrogen and fasting C-peptide.

4. The method for predicting type 2 diabetes based on medical data processing according to claim 1, characterized in that: In S2, the first data preprocessing operation is: sample screening and feature encoding.

5. The method for predicting type 2 diabetes based on medical data processing according to claim 1, characterized in that: In S3.3.1, the similarity comparison is specifically as follows: first, a sample with the same gender as the sample to be supplemented and an age difference of less than 5 years is obtained as a comparison sample, and the similarity between the laboratory indicator values ​​of the sample to be supplemented except for the indicator to be supplemented and the laboratory indicator values ​​of the same type in the comparison sample is calculated, and then the average of the multiple similarities is calculated, and the similarity comparison between the sample to be supplemented and other samples is realized based on the above similarity average value.

6. The method for predicting type 2 diabetes based on medical data processing according to claim 1, characterized in that: The S3.3.2 is specifically: calculating the average value of the values ​​of the indicator to be supplemented in other samples with a similarity exceeding 80% as the value of the indicator to be supplemented in the sample to be supplemented.

7. The method for predicting type 2 diabetes based on medical data processing according to claim 1, characterized in that: In S3.3.3, the second supplementation method is to use a deep learning model to perform data correction operations on the samples to be supplemented.

8. The method for predicting type 2 diabetes based on medical data processing according to claim 7, characterized in that: The deep learning model is a BP neural network model.

9. The method for predicting type 2 diabetes based on medical data processing according to claim 8, characterized in that: The formula for determining the number M of neurons in the hidden layer of the BP neural network model is: Where M s is the number of sample sets of the BP neural network model, M o is the number of neurons in the output layer of the BP neural network model, M i is the number of neurons in the input layer of the BP neural network model, and a is the adjustment coefficient, which is within the range of 3-5.

10. A type 2 diabetes prediction system based on medical data processing, characterized in that: The system adopts a method for predicting type 2 diabetes based on medical data processing according to any one of claims 1 to 9, and the system further comprises: A sample set acquisition module, which acquires a sample set for type 2 diabetes prediction; A first preprocessing module, configured to perform a first data preprocessing operation on the sample set to obtain a second sample set; A second preprocessing module, configured to perform a second data preprocessing operation on the second sample set to obtain a third sample set; A quality detection module, used for performing quality detection on the samples in the third sample set that have undergone the supplementation operation, to obtain a fourth sample set; The type 2 diabetes prediction model is used to retrain the type 2 diabetes prediction model using the fourth sample set, input the patient data to be predicted into the trained type 2 diabetes prediction model to obtain a prediction result.

Images