A device for guiding theophylline medication based on weighted regression

Abstract

The present invention relates to a theophylline medication guidance device based on weighted regression, comprising: a data set acquisition module: used to acquire a data set, wherein the data set is theophylline concentration in the serum collected at different times after different patients take different doses of theophylline; a classification module: used to divide the data set into a training set and a test set; a model construction module: used to construct a regression model, train the training set using the regression model to obtain a trained regression model, and test the trained regression model using the test set to obtain a fitted regression model; a prediction module: used to predict the serum theophylline concentration of patients who need medication guidance using the fitted regression model. The present invention can provide accurate theophylline medication guidance to patients.

Description

Technical Field

[0001] The invention relates to the technical field of auxiliary medical treatment, in particular to a theophylline medication guidance device based on weighted regression. Background Art

[0002] Theophylline can reduce smooth muscle tension and dilate the airways; it can promote the release of endogenous adrenaline and norepinephrine, relax airway smooth muscles; inhibit the release of calcium ions from the endoplasmic reticulum of smooth muscles, reduce intracellular calcium ion concentration and produce airway dilation. Theophylline has a strong relaxant effect on smooth muscles.

[0003] Theophylline is a commonly used drug for the treatment of bronchial and cardiac asthma. The traditional medication regimen is based on the doctor's experience to provide qualitative guidance to achieve the purpose of treatment. It is difficult to quantitatively guide medication according to the situation of each patient. The method of human judgment is often inaccurate. If the dosage is too large, it will have a negative impact on the treatment of cardiac asthma. If the dosage is too small, it will not achieve the purpose of treating cardiac asthma.

[0004] In order to control the onset of chronic asthma in clinical patients, the serum theophylline concentration must reach 8-20μg / ml. Exceeding 20μg / ml can cause poisoning reactions. However, there are many factors affecting the serum theophylline concentration, and there are large individual differences. In actual medication scenarios, it is difficult to personalize and quantitatively administer the drug according to the patient's clinical manifestations, resulting in irreversible consequences.

[0005] Therefore, it is necessary to find a suitable method to provide theophylline suitable for the condition of patients with cardiac asthma, so as to achieve a therapeutic effect. Summary of the invention

[0006] The technical problem to be solved by the present invention is to provide a theophylline medication guidance device based on weighted regression, which can provide accurate theophylline medication guidance for patients.

[0007] The technical solution adopted by the present invention to solve the technical problem is: to provide a theophylline medication guidance device based on weighted regression, comprising:

[0008] Data set acquisition module: used to acquire data sets, wherein the data sets are serum theophylline concentrations collected at different times after different patients take different doses of theophylline;

[0009] Classification module: used to divide the data set into a training set and a test set;

[0010] Model building module: used to build a regression model, train the training set using the regression model to obtain a trained regression model, and test the trained regression model using the test set to obtain a fitted regression model;

[0011] Prediction module: used to predict the serum theophylline concentration of patients who need medication guidance through the fitted regression model.

[0012] The classification module divides the data set into a training set and a test set, specifically: the data set is divided into 80% training set and 20% test set.

[0013] The regression model in the model building module is a weighted linear regression model.

[0014] The model building module constructs a regression model, and uses the regression model to train a training set to obtain a trained regression model, and also includes: when using the regression model to train the training set, assigning different weights to data points with different offset values ​​in the training set, wherein the offset value is the distance between the data point in the training set and the line or plane corresponding to the regression model.

[0015] Different weights are given to data points with different offset values ​​in the training set, and the formula is: Where ω represents the i-th data point x in the training set i , M represents the number of data points in the training set, μ represents the mean of the M data points, S represents the covariance matrix of the M data points, and k represents a constant.

[0016] The model building module tests the trained regression model with the test set to obtain a fitted regression model, specifically: using the test set to test the trained regression model to obtain a test result, and comparing the test result with the true result to calculate the determination coefficient R^2 of the trained regression model;

[0017] If the determination coefficient R^2 is greater than a preset value, the trained regression model is used as a fitted regression model, and the fitted regression model is output;

[0018] If the determination coefficient R^2 is less than a preset value, the weight of the trained regression model is adjusted until the determination coefficient R^2 is greater than the preset value.

[0019] The preset value is 0.7.

[0020] Beneficial Effects

[0021] Due to the adoption of the above-mentioned technical scheme, the present invention has the following advantages and positive effects compared with the prior art: compared with the traditional linear regression method, the weighted linear regression estimated by the present invention has a smaller standard error of the regression coefficient (more stable, less variable) and better model fitting; the present invention combines traditional medicine with modern computer technology to achieve accurate and personalized medication guidance plans; patients of the present invention no longer need to go to the hospital to obtain guidance on medication plans, thereby reducing medical costs and improving treatment efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 is a flow chart of an embodiment of the present invention. DETAILED DESCRIPTION

[0023] The present invention will be further described below in conjunction with specific embodiments. It should be understood that these embodiments are only used to illustrate the present invention and are not intended to limit the scope of the present invention. In addition, it should be understood that after reading the content taught by the present invention, those skilled in the art can make various changes or modifications to the present invention, and these equivalent forms fall within the scope limited by the appended claims of the application equally.

[0024] The embodiment of the present invention relates to a theophylline medication guidance device based on weighted regression, please refer to Figure 1 ,include:

[0025] The data set acquisition module is used to acquire a data set, wherein the data set is theophylline concentration in serum collected at different times after different patients take different doses of theophylline.

[0026] Classification module: used to divide the data set into a training set and a test set.

[0027] The classification module divides the data set into a training set and a test set, specifically: the data set is divided into 80% training set and 20% test set.

[0028] Model building module: used to build a regression model, use the regression model to train the training set to obtain a trained regression model, and test the trained regression model with the test set to obtain a fitted regression model.

[0029] Furthermore, the regression model in the model building module is a weighted linear regression model.

[0030] Traditional linear regression uses ordinary least squares to fit the regression model to obtain the residual, and regresses the appropriate explanatory variables with the square of the residual (or the absolute value of the residual) to obtain the regression equation. However, the unified treatment of data points with different offset values ​​will affect the fitting effect of the model. Therefore, this implementation method improves the data points with different offset values, as follows:

[0031] Furthermore, the construction of the regression model in the model construction module, using the regression model to train the training set to obtain a trained regression model, also includes: when using the regression model to train the training set, assigning different weights to data points with different offset values ​​in the training set, wherein the offset value is the distance between the data point in the training set and the line or plane corresponding to the regression model. The line or plane here (the corresponding line in two-dimensional space and the corresponding plane in multidimensional space) is the line or plane when solving the regression problem, that is, a visualization result of the regression model in this embodiment.

[0032] Different weights are given to data points with different offset values ​​in the training set, and the formula is: Where ω represents the i-th data point x in the training set i , M represents the number of data points in the training set, μ represents the mean of the M data points, S represents the covariance matrix of the M data points, and k represents a constant.

[0033] Furthermore, the model building module tests the trained regression model with the test set to obtain a fitted regression model, specifically: testing the trained regression model with the test set to obtain a test result, comparing the test result with the true result to calculate the determination coefficient R^2 of the trained regression model;

[0034] If the determination coefficient R^2 is greater than a preset value, the trained regression model is used as a fitted regression model, and the fitted regression model is output;

[0035] If the determination coefficient R^2 is less than a preset value, the weight of the trained regression model is adjusted until the determination coefficient R^2 is greater than a preset value. The preset value is 0.7.

[0036] Prediction module: used to predict the serum theophylline concentration of patients who need medication guidance through the fitted regression model. When the theophylline concentration is less than 8μg / ml, it means that the patient needs to take medicine. When the theophylline concentration is between 8 and 20μg / ml, it means that the theophylline is exerting its effect and no medicine is needed. When the theophylline concentration is greater than 20μg / ml, it means that the patient has taken too much medication or the medication interval is too short, there is a risk of poisoning, and the patient should seek medical attention for further examination in time.

[0037] The foregoing description of specific exemplary embodiments of the present invention is for the purpose of illustration and demonstration. These descriptions are not intended to limit the present invention to the precise form disclosed, and it is clear that many changes and variations can be made based on the above teachings. The purpose of selecting and describing the exemplary embodiments is to explain the specific principles of the present invention and its practical application, so that those skilled in the art can realize and utilize various different exemplary embodiments of the present invention and various different selections and changes. The scope of the present invention is intended to be limited by the claims and their equivalents.

Claims

1. A theophylline medication guidance device based on weighted regression, characterized in that: include: Data set acquisition module: used to acquire data sets, wherein the data sets are serum theophylline concentrations collected at different times after different patients take different doses of theophylline; Classification module: used to divide the data set into a training set and a test set; Model building module: used to build a regression model, use the regression model to train the training set to obtain a trained regression model, and test the trained regression model with the test set to obtain a fitted regression model; wherein the regression model in the model building module is a weighted linear regression model, and when the training set is trained with the regression model, different weights are assigned to data points with different offset values ​​in the training set, wherein the offset value is the distance between the data point in the training set and the line or plane corresponding to the regression model; Different weights are given to data points with different offset values ​​in the training set, and the formula is: Where ω represents the i-th data point x in the training set i , M represents the number of data points in the training set, μ represents the mean of the M data points, S represents the covariance matrix of the M data points, and k represents a constant; The model building module tests the trained regression model using the test set to obtain a fitted regression model, specifically: The trained regression model is tested using the test set to obtain a test result, and the test result is compared with the true result to calculate the coefficient of determination R^2 of the trained regression model; If the determination coefficient R^2 is greater than a preset value, the trained regression model is used as a fitted regression model, and the fitted regression model is output; If the determination coefficient R^2 is less than a preset value, the weight of the trained regression model is adjusted until the determination coefficient R^2 is greater than the preset value; Prediction module: used to predict the serum theophylline concentration of patients who need medication guidance through the fitted regression model.

2. The theophylline medication guidance device based on weighted regression according to claim 1, characterized in that: The classification module divides the data set into a training set and a test set, specifically: the data set is divided into 80% training set and 20% test set.

3. The theophylline medication guidance device based on weighted regression according to claim 1, characterized in that: The preset value is 0.7.

Images