Patient data preprocessing method and device and electronic equipment
By classifying and preprocessing patient data and using recursive estimation and missing value filling model, the problem of poor quality of wearable sensor data is solved, the accuracy and consistency of data is improved, and high-quality basic data is provided for treatment plans.
Patent Information
- Application Number
- CN202410018303.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-01-05
- Publication Date
- 2025-07-08
AI Technical Summary
In the prior art, patient data is less quality due to the poor data measured by wearable sensors, and is affected by factors such as electromagnetic interference, temperature, humidity and human operation errors, resulting in more data noise, affecting the accuracy of the treatment plan.
By classifying patient data, using recursive estimation processing and missing value filling model, predicted state data is generated on the data collected by the wearable sensor, and missing value filling data collected by the application are filled to improve data quality.
It improves the accuracy and consistency of patient data, ensures the targetedness and accuracy of subsequent treatment plans, and is suitable for dynamically changing non-steady state data.
Smart Images
Figure CN120280062A_ABST
Abstract
Description
Technical Field
[0001] The embodiments of the present application relate to the field of medical technologies, and particularly to a method, an apparatus, and an electronic device for preprocessing patient data. Background Art
[0002] After a user wears a wearable sensor, patient data is collected through the wearable sensor. In addition, patient data can also be collected based on application programs installed on some other wearable portable devices. Furthermore, these patient data can be used for remote medical diagnosis.
[0003] However, directly using these patient data to match treatment plans, due to the fact that the data measured by wearable sensors often has a lot of noise, such as electromagnetic interference, being affected by temperature and humidity, or human operation errors, etc., the quality of the data is thus poor. Summary of the Invention
[0004] The purpose of the present application is to propose a method, an apparatus, and an electronic device for preprocessing patient data to solve or overcome the above-mentioned technical problems existing in the prior art.
[0005] According to the first aspect of the embodiments of the present application, a method for preprocessing patient data is provided, which includes:
[0006] Obtain patient data collected by a patient data collection module;
[0007] Classify the patient data to obtain first-type data and second-type data, where the first-type data is collected by a wearable sensor configured on a patient, and the second-type data is collected by an application program;
[0008] Perform recursive estimation processing on the first-type data based on a set first preprocessing model to generate corresponding predicted state data;
[0009] Perform missing value filling on the second-type data based on a set second preprocessing model to generate corresponding estimated missing values.
[0010] Optionally, the following steps are performed based on the set first preprocessing model to perform recursive estimation processing on the first-type data to generate corresponding predicted state data:
[0011] Obtain the historical estimated value of the first-type data;
[0012] Based on the historical estimated value, perform recursive estimation processing on the first-type data to generate corresponding predicted state data.
[0013] Optionally, recursively estimating the first type of data based on the current measurement value and the historical estimate to generate corresponding predicted state data, including:
[0014] Determining a current estimate based on a set state transition matrix and the historical estimate;
[0015] Recursively estimating the first type of data based on the current estimate and the Kalman gain calculated according to the state transition matrix to generate corresponding predicted state data.
[0016] Optionally, the method further includes:
[0017] Determining a predicted error covariance at the current moment based on the state transition matrix and the historical predicted error covariance;
[0018] Calculating the Kalman gain according to the predicted error covariance at the current moment and the state transition matrix.
[0019] Optionally, the method further includes:
[0020] Dividing the predicted state data into a number of data samples according to a set time period;
[0021] Performing a short-time Fourier transform on each data sample to generate a short-time spectrum;
[0022] Generating a micro-Doppler time-frequency map according to the short-time spectra corresponding to all samples.
[0023] Optionally, filling in the missing values of the second type of data based on a set second preprocessing model to generate corresponding estimated missing values, including:
[0024] Assigning a separate category to the second type of data with missing values;
[0025] Predicting the missing values for the category based on the set second preprocessing model to generate corresponding estimated missing values.
[0026] Optionally, the method further includes: determining whether the second type of data with missing values is a categorical variable or a continuous variable;
[0027] If the second type of data with missing values is a categorical variable, performing approximate sample prediction on the category based on the set second preprocessing model to generate corresponding estimated missing values;
[0028] If the second type of data with missing values is a continuous variable, performing approximate sample prediction or a statistic filling algorithm on the category based on the set second preprocessing model to generate corresponding estimated missing values.
[0029] Optionally, the method further includes:
[0030] Performing one-hot encoding on the categorical variable to convert it into a binary variable; and / or
[0031] Performing linearization transformation on the continuous variable to obtain linear data.
[0032] According to a second aspect of the embodiments of the present application, a preprocessing device for patient data is provided, including:
[0033] A data acquisition unit, configured to acquire patient data collected by a patient data collection module;
[0034] A classification unit, configured to classify the patient data to obtain first-type data and second-type data, where the first-type data is collected by a wearable sensor configured on a patient, and the second-type data is collected by an application;
[0035] A first prediction unit, configured to perform recursive estimation processing on the first-type data based on a set first preprocessing model to generate corresponding predicted state data;
[0036] A second prediction unit, configured to fill in missing values for the second-type data based on a set second preprocessing model to generate corresponding estimated missing values.
[0037] According to a third aspect of the embodiments of the present application, an electronic device is provided, including a memory and a processor, where a computer-executable program is stored on the memory, and the processor is configured to run the computer-executable program to execute the method according to any one of the embodiments of the present application. Description of the Drawings
[0038] Some specific embodiments of the embodiments of the present application will be described in detail hereinafter with reference to the drawings in an exemplary and non-limiting manner. The same reference numerals in the drawings denote the same or similar components or parts. Those skilled in the art should understand that these drawings are not necessarily drawn to scale. In the drawings:
[0039] Figure 1 It is a schematic flowchart of a method for preprocessing patient data according to an embodiment of the present application.
[0040] Figure 2 It is a schematic flowchart of step S103 according to an embodiment of the present application.
[0041] Figure 3 It is a schematic flowchart of step S104 according to an embodiment of the present application.
[0042] Figure 4 It is a schematic structural diagram of a preprocessing device for patient data according to an embodiment of the present application. Specific implementation mode
[0043] In order to enable those skilled in the art to better understand the technical solutions in the embodiments of the present application, the following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the embodiments of the present application, all other embodiments obtained by those of ordinary skill in the art shall fall within the scope of protection of the embodiments of the present application.
[0044] Figure 1 It is a schematic flowchart of a method for preprocessing patient data in an embodiment of the present application. As Figure 1 shown, it includes:
[0045] S101. Obtain the patient data collected by the patient data collection module;
[0046] S102. Classify the patient data to obtain first-type data and second-type data. The first-type data is collected by wearable sensors configured on the patient, and the second-type data is collected by an application program;
[0047] S103. Perform recursive estimation processing on the first-type data based on the set first preprocessing model to generate corresponding predicted state data;
[0048] S104. Fill in the missing values of the second-type data based on the set second preprocessing model to generate corresponding estimated missing values.
[0049] In the above embodiment, after obtaining the patient data collected by the patient data collection module, the patient data is classified to obtain first-type data and second-type data, so as to distinguish which are collected by wearable sensors and which are collected by the application program, realizing the preliminary classification of patient data. Further, for these two types of data, targeted processing is performed based on different preprocessing models respectively, realizing the deep-level preprocessing of data, improving the quality of patient data, and thus ensuring that the subsequent treatment plan generated has high accuracy and can achieve targeted treatment of patients.
[0050] Optionally, Figure 2 It is a schematic flowchart of step S103 in an embodiment of the present application. As Figure 2 shown, in step S103, the following steps are performed based on the set first preprocessing model to perform recursive estimation processing on the first-type data to generate corresponding predicted state data:
[0051] S113. Obtain the historical estimated value of the first-type data;
[0052] S123. Recursively estimate the first type of data based on the historical estimate to generate corresponding predicted state data.
[0053] Optionally, in step S123, recursively estimating the first type of data based on the current measurement value and the historical estimate to generate corresponding predicted state data includes:
[0054] S1231. Determine the current estimate based on a set state transition matrix and the historical estimate;
[0055] S1232. Recursively estimate the first type of data based on the current estimate and the Kalman gain calculated according to the state transition matrix to generate corresponding predicted state data.
[0056] Optionally, the method further includes:
[0057] Determine the predicted error covariance at the current moment based on the state transition matrix and the historical predicted error covariance;
[0058] Calculate the Kalman gain according to the predicted error covariance at the current moment and the state transition matrix.
[0059] In a specific application scenario, the recursive estimation process of the first type of data based on a set first preprocessing model to generate corresponding predicted state data can be implemented according to the following formula (1):
[0060]
[0061] where, is the preliminary predicted state data at time k, i.e., the current estimate, A is the state transition matrix, is the estimated state at time k - 1, i.e., the historical estimate, and k is a positive integer.
[0062] Thus, the current estimate is determined based on the set state transition matrix and the historical estimate through the above formula (1).
[0063]
[0064] where, is the predicted error covariance at time k, i.e., the predicted error covariance at the current moment, P k-1 is the error covariance at time k - 1, i.e., the historical predicted error covariance, and Q is the process noise covariance matrix. That is, the predicted error covariance at the current moment is determined based on the state transition matrix and the historical predicted error covariance through the above formula (2).
[0065]
[0066] K k is the Kalman gain at time k, H is the observation matrix, and R is the measurement noise covariance matrix.
[0067] That is, according to the above formula (3), the Kalman gain is calculated based on the prediction error covariance at the current time and the state transition matrix.
[0068]
[0069] is the final state estimate at time k, that is, the predicted state data corresponding to the first type of data at time k, z k is the current measurement value of the first type of data.
[0070] On the basis of the above processing, in order to facilitate the calculation of the predicted state data corresponding to the first type of data at time k + 1, it further includes: updating the prediction error covariance at the current time according to the prediction error covariance at the current time and the observation matrix. For example, it is implemented with reference to the following formula (5).
[0071]
[0072] P k is the final error covariance at time k, and I is the identity matrix.
[0073] In the above embodiments, each matrix can be determined according to the actual application scenario. The process noise covariance Q, the measurement noise covariance matrix R, and the prediction error covariance matrix P are set parameters. First, an initial value is given by experience, and then it is continuously optimized according to the results. The observation matrix H is directly obtained by observation. Both the Kalman gain K and the state transition matrix A can be obtained through formula derivation.
[0074] In this embodiment, based on the above first preprocessing model, it is possible to accurately predict the future state of the data on the basis of considering the current measurement value and the historical estimated value, so that it can be applied to the non-steady first type of data and can also adapt to the dynamically changing first type of data, thereby providing a more accurate state estimate and providing more stable data for subsequent analysis.
[0075] The wearable sensors that generate the above first type of data include, but are not limited to, temperature sensors, humidity sensors, etc.
[0076] On the basis of the above embodiments, optionally, the method further includes:
[0077] Dividing the predicted state data into several data samples according to a set time period;
[0078] Perform short-time Fourier transform on each data sample to generate a short-time spectrum;
[0079] Generate based on the short-time spectra corresponding to all samples.
[0080] Optionally, in this embodiment, for example, the predicted state data can be divided according to a set time window function so as to divide the predicted state data according to a set time period, and then several data samples are obtained.
[0081] Optionally, a micro-Doppler time-frequency map is generated based on the short-time spectra corresponding to all samples. For example, the short-time spectra within all time window functions can be stacked to obtain a micro-Doppler time-frequency map, so as to visually observe the changes of the predicted state data in time and frequency, thereby providing high-quality basic data for subsequent training and testing.
[0082] In addition, since the micro-Doppler time-frequency map is formed by the above stacking, the matrixization of the predicted state data is realized, and the predicted state data is embodied in the form of a matrix, thereby providing high-quality basic data for subsequent training and testing.
[0083] Optionally, Figure 3 is a schematic flow chart of step S104 of the embodiment of the present application. As Figure 3 shown, in the step S104, based on a set second preprocessing model, missing value filling is performed on the second type of data to generate corresponding estimated missing values, including:
[0084] S114. Assign an independent category to the second type of data with missing values;
[0085] S124. Perform missing value prediction on the category based on the set second preprocessing model to generate corresponding estimated missing values.
[0086] The above second type of data is, for example, the basic attribute data of a patient, which may include gender, region, education level, age, etc.
[0087] In this embodiment, by assigning the above-mentioned opposing categories, the information of the missing values in the second type of data can be retained, and no additional bias or distortion is introduced. Further, excessive processing of the data can be avoided, thereby maintaining the integrity and consistency of the data.
[0088] Optionally, the method further includes: determining whether the second type of data with missing values is a categorical variable or a continuous variable. For example, the gender is a categorical variable, while the age is a continuous variable.
[0089] If the second type of data with missing values is a categorical variable, perform approximate sample prediction on the category based on the set second preprocessing model to generate corresponding estimated missing values;
[0090] If the second type of data with missing values is a continuous variable, perform approximate sample prediction or statistical filling algorithm on the category based on the set second preprocessing model to generate corresponding estimated missing values.
[0091] Optionally, the approximate sample prediction on the category based on the set second preprocessing model in the above steps can be implemented based on the following steps:
[0092]
[0093] Where: is the predicted category of the missing value x, c j is a category in the set of optional categories, K is the number of selected nearest neighbors, y i is the category of the i-th training sample closest to the missing value x, I(·) is the indicator function, if the condition in the parentheses is true, the result is 1, otherwise it is 0.
[0094] Furthermore, assuming the estimated missing value is x and the training data set is D, It can be expressed by the following formula:
[0095]
[0096] Where, is the estimated missing value of the missing value x, K is the number of selected nearest neighbors, y i is the value of the i-th training sample closest to the missing value x.
[0097] In the above embodiments, the training samples are constructed according to the usage scenario to achieve the estimation of the above missing values.
[0098] The above formulas (6)(7) can be applied to both the case where the second type of data with missing values is a categorical variable and the case of a continuous variable.
[0099] Optionally, the method further includes:
[0100] Perform one-hot encoding on the categorical variable to convert it into a binary variable; and / or
[0101] Perform linearization transformation on the continuous variable to obtain linear data.
[0102] In this embodiment, through the above one-hot encoding, the reasonable expression of data in the model can be ensured, thereby improving the expression ability of data in the model, reducing the scale difference between features, improving the performance of the model and enhancing the stability of the model when applied to model training.
[0103] In the binary variables, some variables take the value of 1 and some variables take the value of 0. This encoding method can effectively handle the magnitude relationship between categorical variables.
[0104] Taking the one-hot encoding of the color column that can be used as a categorical variable as an example:
[0105] Before encoding:
[0106] … Color … Red Green Red
[0107] After encoding:
[0108]
[0109]
[0110] In this embodiment, when linearizing the continuous variables, for example, it can be achieved through techniques such as logarithmic transformation and normalization to ensure the reasonable expression of data in the model, thereby improving the expression ability of data in the model, reducing the scale difference between features, and improving the performance and stability of the model when applied to model training.
[0111] Optionally, considering that continuous variables may exhibit a right-skewed (long-tailed) or left-skewed (short-tailed) distribution, therefore, when linearizing the continuous variables to obtain linear data through logarithmic transformation, specifically, the power relationship of the original data of the continuous variables is transformed into a linear relationship, better presenting the distribution characteristics of the data, making the transformed data closer to a normal distribution, and reducing the impact of skewness such as right-skewed (long-tailed) or left-skewed (short-tailed) on model training.
[0112] In this embodiment, if the linearization of the continuous variables is achieved through normalization, data with different features or different data ranges can be transformed into a unified scale, so that the model can better process and learn, improving the performance and convergence speed of the model.
[0113] Exemplarily, mean normalization can be any one of the following three.
[0114] 1. Min-Max normalization (Min-Max Scaling):
[0115]
[0116] Where X normalizedis the normalized value, X is the original data of the continuous variable, X min and X max are the minimum and maximum values of the continuous variable respectively.
[0117] 2. Z-Score Normalization (Standardization)
[0118] By zero-centering and unit-variance normalizing the continuous variable, the distribution of the data is made to approximate the standard normal distribution (mean = 0, standard deviation = 1).
[0119]
[0120] where, X normalized is the normalized value, X is the original data of the continuous variable, μ is the mean of the data, and σ is the standard deviation of the data.
[0121] 3. Range Scaling Method (Scaling to Unit Length):
[0122] The data is scaled to unit length in vector form, considering the direction of the eigenvector rather than just the magnitude, which is especially applicable to the text data of patients.
[0123]
[0124] where, X normalized is the normalized vector, X is the original data vector of the continuous variable, and ‖X‖ represents the L2 norm of the original data vector of the continuous variable.
[0125] Alternatively, if the second type of data with missing values is a continuous variable, in addition to being able to perform approximate sample prediction or statistical filling algorithms on the category based on the set second preprocessing model to generate corresponding estimated missing values, it is also possible to convert the continuous variable into a categorical variable based on a classification machine learning algorithm such as a decision tree model, and then generate an estimated loss value based on the missing value estimation method for categorical variables provided above.
[0126] When converting a continuous variable into a categorical variable, the method of information gain entropy is used to calculate an information gain entropy as the initial splitting point to divide the continuous variable into two subsets, and then the information entropy of each subset is calculated. Finally, the best threshold is selected through information gain, and the patients are classified based on this best threshold, such as into a high-risk group and a low-risk group, thus implementing a dichotomy to further judge the logical relationship between the high-risk group and the low-risk group, such as the case where the risk of disease is significantly higher when cholesterol is greater than a certain value than when it is less than a certain value (i.e., high cholesterol accelerates the deterioration of the disease).
[0127] Entropy measures the degree of disorder of a dataset, and its calculation formula is as follows:
[0128]
[0129] Among them, S is the dataset composed of continuous variables, c is the number of categories of continuous variables, and p i is the proportion of category i in the dataset.
[0130] Information Gain is the criterion for selecting the best splitting point in a decision tree, and its calculation formula is as follows:
[0131]
[0132] Among them, S is the dataset of the parent node, A is a continuous variable, Values(A) is the subset divided according to the value of the continuous variable A, and S v is the subset where the value of the continuous variable A is v, and |S v | and |S| are the number of samples of the subset S v and the dataset S of the parent node respectively.
[0133] The steps are as follows:
[0134] 1. Sort the continuous variable A from small to large.
[0135] 2. Select the average of two adjacent values as the threshold and calculate the information gain.
[0136] 3. Repeat step 2 until the information gain of all possible thresholds is calculated.
[0137] 4. Select the threshold with the maximum information gain as the classification criterion and divide the continuous variable into two classifications.
[0138] Figure 4 is the structural schematic diagram of the preprocessing device for patient data in the embodiment of the present application. As Figure 4 shown, it includes:
[0139] A data acquisition unit 401, configured to acquire patient data collected by a patient data collection module;
[0140] A classification unit 402, configured to classify the patient data to obtain first-type data and second-type data, where the first-type data is collected by wearable sensors configured on the patient, and the second-type data is collected by an application;
[0141] A first prediction unit 403, configured to perform recursive estimation processing on the first-type data based on a set first preprocessing model to generate corresponding predicted state data;
[0142] A second prediction unit 404, configured to perform missing value filling on the second type of data based on a set second preprocessing model to generate corresponding estimated missing values.
[0143] Optionally, when the first prediction unit 403 performs the following steps based on a set first preprocessing model to perform recursive estimation processing on the first type of data to generate corresponding prediction status data:
[0144] Obtain historical estimated values of the first type of data;
[0145] Based on the historical estimated values, perform recursive estimation processing on the first type of data to generate corresponding prediction status data.
[0146] Optionally, when the first prediction unit 403 performs recursive estimation processing on the first type of data based on the current measurement value and the historical estimated values to generate corresponding prediction status data, the following steps are performed:
[0147] Determine a current estimated value based on a set state transition matrix and the historical estimated values;
[0148] Based on the current estimated value and the Kalman gain calculated according to the state transition matrix, perform recursive estimation processing on the first type of data to generate corresponding prediction status data.
[0149] Optionally, the device further includes:
[0150] A variance calculation unit, configured to determine a current moment prediction error covariance based on the state transition matrix and a historical prediction error covariance;
[0151] A gain calculation unit, configured to calculate the Kalman gain according to the current moment prediction error covariance and the state transition matrix.
[0152] Optionally, the device further includes:
[0153] A partitioning unit, configured to partition the prediction status data according to a set time period to obtain a plurality of data samples;
[0154] A transformation unit, configured to perform a short-time Fourier transform on each data sample to generate a short-time spectrum;
[0155] A time-frequency diagram unit, configured to generate a micro-Doppler time-frequency diagram according to the short-time spectra corresponding to all samples.
[0156] Optionally, when the second prediction unit 404 performs missing value filling on the second type of data based on a set second preprocessing model to generate corresponding estimated missing values, the following steps are performed:
[0157] Assign a separate category to the second type of data with missing values;
[0158] Based on the set second preprocessing model, predict the missing values for the category to generate corresponding estimated missing values.
[0159] Optionally, the device further includes: a variable judgment unit, configured to judge whether the second type of data with missing values is a categorical variable or a continuous variable;
[0160] If the second type of data with missing values is a categorical variable, the second prediction unit 404 performs approximate sample prediction on the category based on the set second preprocessing model to generate corresponding estimated missing values;
[0161] If the second type of data with missing values is a continuous variable, the second prediction unit 404 performs approximate sample prediction or a statistic filling algorithm on the category based on the set second preprocessing model to generate corresponding estimated missing values.
[0162] Optionally, the device further includes a data post-processing unit, configured to perform one-hot encoding on the categorical variable to convert it into a binary variable; and / or perform a linearization transformation on the continuous variable to obtain linear data.
[0163] According to the third aspect of the embodiments of the present application, an electronic device is provided, which includes a memory and a processor, and a computer-executable program is stored on the memory, and the processor is configured to run the computer-executable program to execute the method according to any one of the embodiments of the present application.
[0164] The above description is only a preferred embodiment of the present application and an explanation of the applied technical principle. Those skilled in the art should understand that the scope of the invention involved in the present application is not limited to the technical solution formed by the specific combination of the above technical features, and should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above inventive concept. For example, a technical solution formed by mutually replacing the above features with (but not limited to) technical features with similar functions disclosed in the present application.
Claims
1. A method for preprocessing patient data, characterized in that including: Obtaining patient data collected by a patient data collection module; Classifying the patient data to obtain first-type data and second-type data, where the first-type data is collected by wearable sensors configured on the patient, and the second-type data is collected by an application; Performing recursive estimation processing on the first-type data based on a set first preprocessing model to generate corresponding predicted state data; Performing missing value filling on the second-type data based on a set second preprocessing model to generate corresponding estimated missing values.
2. The method according to claim 1, wherein The method based on the set first preprocessing model performs the following steps to perform recursive estimation processing on the first-type data to generate corresponding predicted state data: Obtaining historical estimated values of the first-type data; Based on the historical estimated values, performing recursive estimation processing on the first-type data to generate corresponding predicted state data.
3. The method according to claim 2, wherein The performing recursive estimation processing on the first-type data based on the current measured value and the historical estimated values to generate corresponding predicted state data includes: Determining a current estimated value based on a set state transition matrix and the historical estimated values; Based on the current estimated value and the Kalman gain calculated according to the state transition matrix, performing recursive estimation processing on the first-type data to generate corresponding predicted state data.
4. The method according to claim 3, wherein The method further includes: Determining a current-time prediction error covariance based on the state transition matrix and the historical prediction error covariance; Calculating the Kalman gain according to the current-time prediction error covariance and the state transition matrix.
5. The method according to any one of claims 1-4, characterized in that, The method further includes: Dividing the predicted state data according to a set time period to obtain a number of data samples; Performing short-time Fourier transform on each data sample to generate a short-time spectrum; Generating a micro-Doppler time-frequency diagram according to the short-time spectra corresponding to all samples.
6. The method according to claim 5, wherein The performing missing value filling on the second-type data based on the set second preprocessing model to generate corresponding estimated missing values includes: Assigning a separate category to the second-type data with missing values; Performing missing value prediction on the category based on the set second preprocessing model to generate corresponding estimated missing values.
7. The method according to claim 6, characterized in that The method further includes: determining whether the second-type data with missing values is a categorical variable or a continuous variable; If the second-type data with missing values is a categorical variable, then performing approximate sample prediction on the category based on the set second preprocessing model to generate corresponding estimated missing values; If the second-type data with missing values is a continuous variable, then performing approximate sample prediction or a statistical quantity filling algorithm on the category based on the set second preprocessing model to generate corresponding estimated missing values.
8. The method according to claim 7, wherein The method further includes: Performing one-hot encoding on the categorical variable to convert it into a binary variable; and / or Performing linearization transformation on the continuous variable to obtain linear data.
9. A preprocessing device for patient data, characterized in that, including: A data acquisition unit for obtaining patient data collected by a patient data collection module; A classification unit for classifying the patient data to obtain first-type data and second-type data, where the first-type data is collected by wearable sensors configured on the patient, and the second-type data is collected by an application program; A first prediction unit for performing recursive estimation processing on the first-type data based on a set first preprocessing model to generate corresponding predicted state data; A second prediction unit for filling in missing values of the second-type data based on a set second preprocessing model to generate corresponding estimated missing values.
10. An electronic device, characterized in that, It includes a memory and a processor, and a computer-executable program is stored on the memory, and the processor is used to run the computer-executable program to execute the method according to any one of claims 1-8.