Diagnostic model construction method and device, equipment and medium

By distinguishing and training machine learning models based on meridian impedance data, the problem of insufficient fitting accuracy and adaptability of machine learning algorithm models in the medical field in the prior art is solved, and the accuracy of diagnosis is improved.

CN119092098APending Publication Date: 2024-12-06CAPITALBIO CORP +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411153424.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-21
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

The existing machine learning algorithm models have problems such as low fitting accuracy, low fitting degree and low classification effect accuracy in the application of medical field, resulting in low diagnostic accuracy.

Method used

By obtaining health status data and meridian impedance data, we distinguish between healthy and non-healthy meridian impedance data. If the significant difference is greater than the target level value, we will use the support vector machine algorithm to train the machine learning model and build a diagnostic model.

Benefits of technology

The fitting accuracy, fit and accuracy of the diagnostic model are improved, and the accuracy of the diagnosis is improved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119092098A_ABST
    Figure CN119092098A_ABST
Patent Text Reader

Abstract

The embodiment of the invention provides a construction method and device of a diagnosis model, equipment and a medium, and relates to the technical field of machine learning, and the method comprises the steps: obtaining health condition data and acupuncture point impedance data, the health condition data and the acupuncture point impedance data are in one-to-one correspondence, and the acupuncture point impedance data comprise acupuncture point information and impedance data values; according to the health condition data, dividing the acupuncture point impedance data into healthy acupuncture point impedance data and unhealthy acupuncture point impedance data; if the healthy acupuncture point impedance data and the non-healthy acupuncture point impedance data correspond to the same acupuncture point information, and the significant difference result of the healthy acupuncture point impedance data and the non-healthy acupuncture point impedance data is greater than a target level value, training a machine learning model through the healthy acupuncture point impedance data and the non-healthy acupuncture point impedance data; and obtaining a diagnosis model. Therefore, a diagnosis model with better algorithm model fitting precision, adaptability, accuracy and quantification effect can be obtained, and the diagnosis accuracy is further improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of machine learning technology, and in particular to a method, device, equipment and medium for constructing a diagnostic model. Background Art

[0002] With the rapid development of science and technology, artificial intelligence technology has made significant breakthroughs in finance, education, transportation, medical care and other fields. Especially in the medical field, machine learning technology, as the core of artificial intelligence technology, has been widely used in image analysis, disease diagnosis, treatment effect evaluation and other aspects of the medical field.

[0003] In related technologies, machine learning algorithms can automatically extract valuable features and patterns through deep learning and analysis of large amounts of medical data, assisting doctors in making accurate disease diagnosis and evaluating treatment effects. This not only greatly improves the efficiency of medical services, but also makes the diagnosis results more objective and accurate, providing patients with a better medical experience.

[0004] However, the algorithm models of existing machine learning technologies are relatively fixed, and many medical applications directly apply existing mainstream machine learning algorithms and parameters. This fixed approach limits the flexibility and adaptability of machine learning algorithms. When faced with complex and changing medical data and scenarios, the algorithm model is prone to low fitting accuracy, low adaptability, low classification accuracy, and low fit with actual application scenarios, which in turn leads to low diagnostic accuracy. Summary of the invention

[0005] Based on the above problems, the present application provides a method, device, equipment and medium for constructing a diagnostic model, which can improve the accuracy of diagnosis.

[0006] The embodiments of the present application disclose the following technical solutions:

[0007] In a first aspect, the present application discloses a method for constructing a diagnostic model, the method comprising:

[0008] Acquire health condition data and acupoint impedance data, wherein the health condition data and the acupoint impedance data correspond to each other one by one, and the acupoint impedance data includes acupoint information and impedance data values;

[0009] According to the health condition data, the acupoint impedance data is divided into healthy acupoint impedance data and unhealthy acupoint impedance data;

[0010] If the healthy acupoint impedance data and the unhealthy acupoint impedance data correspond to the same acupoint information, and the significant difference between the healthy acupoint impedance data and the unhealthy acupoint impedance data is greater than the target level value, a machine learning model is trained by using the healthy acupoint impedance data and the unhealthy acupoint impedance data to obtain a diagnostic model.

[0011] Optionally, the training of a machine learning model by using the healthy acupoint impedance data and the unhealthy acupoint impedance data includes:

[0012] The healthy acupoint impedance data and the unhealthy acupoint impedance data are divided into a training set and a validation set;

[0013] Select the objective kernel function of the machine learning model;

[0014] Selecting target parameters of the machine learning model by cross-validation method;

[0015] The training set is input into the machine learning model, the kernel function of the machine learning model is the target kernel function, and the parameters of the machine learning model are the target parameters.

[0016] Optionally, the method further comprises:

[0017] Acquire a test set, wherein the test set includes the healthy acupoint impedance data and / or the unhealthy acupoint impedance data;

[0018] Inputting the test set into the machine learning model to obtain a confusion matrix, wherein the confusion matrix includes the accuracy of the machine learning model;

[0019] If the accuracy is lower than the preset accuracy threshold, continue to execute the step of selecting the target kernel function of the machine learning model.

[0020] Optionally, the training of the machine learning model includes:

[0021] Use the support vector machine (SVM) algorithm to train the SVM model.

[0022] Optionally, the method for obtaining the significant difference result between the healthy acupoint impedance data and the unhealthy acupoint impedance data includes:

[0023] By using a wire box diagram method, a visualization operation is performed on the healthy acupoint impedance data and the unhealthy acupoint impedance data to obtain a wire box diagram visualization result;

[0024] By analyzing the box plot visualization result, the significant difference result between the healthy acupoint impedance data and the unhealthy acupoint impedance data is obtained.

[0025] In a second aspect, an embodiment of the present application discloses a diagnostic model construction device, the device comprising: a data acquisition module, a data classification module and a model construction module;

[0026] The data acquisition module is used to acquire health status data and acupoint impedance data, the health status data and the acupoint impedance data correspond to each other one by one, and the acupoint impedance data includes acupoint information and impedance data values;

[0027] The data classification module is used to classify the acupoint impedance data into healthy acupoint impedance data and unhealthy acupoint impedance data according to the health condition data;

[0028] The model building module is used to train a machine learning model through the healthy acupoint impedance data and the unhealthy acupoint impedance data to obtain a diagnostic model if the healthy acupoint impedance data and the unhealthy acupoint impedance data correspond to the same acupoint information and the significant difference result between the healthy acupoint impedance data and the unhealthy acupoint impedance data is greater than a target level value.

[0029] Optionally, the model building module includes: a data differentiation module, a function selection module, a parameter selection module and a data input module;

[0030] The data distinguishing module is used to distinguish the healthy acupoint impedance data and the unhealthy acupoint impedance data into a training set and a validation set;

[0031] The function selection module is used to select the target kernel function of the machine learning model;

[0032] The parameter selection module is used to select the target parameters of the machine learning model through a cross-validation method;

[0033] The data input module is used to input the training set into the machine learning model, the kernel function of the machine learning model is the target kernel function, and the parameters of the machine learning model are the target parameters.

[0034] Optionally, the device further comprises: a test set acquisition module, a matrix acquisition module and a continued selection module;

[0035] The test set acquisition module is used to acquire a test set, wherein the test set includes the healthy acupoint impedance data and / or the unhealthy acupoint impedance data;

[0036] The matrix acquisition module is used to input the test set into the machine learning model to obtain a confusion matrix, wherein the confusion matrix includes the accuracy of the machine learning model;

[0037] The continuing selection module is used to continue executing the step of selecting the target kernel function of the machine learning model if the accuracy is lower than a preset accuracy threshold.

[0038] Optionally, the model training module is specifically used to: train an SVM model using a support vector machine (SVM) algorithm.

[0039] Optionally, the device for acquiring the significant difference result between the healthy acupoint impedance data and the unhealthy acupoint impedance data comprises: a first acquisition module and a second acquisition module;

[0040] The first acquisition module is used to perform a visualization operation on the healthy acupoint impedance data and the unhealthy acupoint impedance data by a wire box diagram method to obtain a wire box diagram visualization result;

[0041] The second acquisition module is used to acquire the significant difference results between the healthy acupoint impedance data and the unhealthy acupoint impedance data by analyzing the wire box diagram visualization results.

[0042] Compared with the prior art, this application has the following beneficial effects:

[0043] The embodiment of the present application provides a method, device, equipment and medium for constructing a diagnostic model, which relates to the field of machine learning technology. The method includes: obtaining health status data and acupoint impedance data, the health status data and acupoint impedance data correspond one to one, and the acupoint impedance data includes acupoint information and impedance data value; according to the health status data, the acupoint impedance data is divided into healthy acupoint impedance data and unhealthy acupoint impedance data; if the healthy acupoint impedance data and the unhealthy acupoint impedance data correspond to the same acupoint information, and the significant difference between the healthy acupoint impedance data and the unhealthy acupoint impedance data is greater than the target level value, then the machine learning model is trained by the healthy acupoint impedance data and the unhealthy acupoint impedance data to obtain a diagnostic model. Thus, a diagnostic model with better algorithm model fitting accuracy, fitness, accuracy and quantification effect can be obtained, thereby improving the accuracy of diagnosis. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0045] Figure 1 A flowchart of a method for constructing a diagnostic model provided in an embodiment of the present application;

[0046] Figure 2 A schematic diagram of a wire box diagram provided in an embodiment of the present application;

[0047] Figure 3 A flowchart of a training SVM model provided in an embodiment of the present application;

[0048] Figure 4 A schematic diagram of a confusion matrix provided in an embodiment of the present application;

[0049] Figure 5 A schematic diagram of a diagnostic model construction device provided in an embodiment of the present application;

[0050] Figure 6 A structural diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0051] First, the technical terms involved in this application are explained:

[0052] Machine learning algorithms are at the core of the field of artificial intelligence, and are designed to acquire new knowledge or skills by simulating or implementing human behavior, reorganizing existing knowledge structures, and continuously improving their own intelligence. Machine learning algorithms can be designed for a given task, with training data as input and a model as output.

[0053] Support Vector Machine (SVM) is a generalized linear classifier that performs binary classification of data in a supervised learning manner. Its decision boundary is the maximum margin hyperplane solved for the learning samples. By introducing the kernel method, SVM can also perform nonlinear classification and is one of the common kernel learning methods.

[0054] As described above, machine learning algorithms can automatically extract valuable features and patterns through deep learning and analysis of large amounts of medical data, assisting doctors in making accurate disease diagnoses and evaluating treatment effects. This not only greatly improves the efficiency of medical services, but also makes diagnostic results more objective and accurate, providing patients with a better medical experience.

[0055] Specifically, in the field of image analysis, machine learning technology can automatically identify and analyze abnormal lesions in medical images, providing doctors with a fast and accurate basis for diagnosis. Through the application of machine learning, doctors can more easily obtain patients' imaging information, conduct quantitative analysis and comparison, and thus more accurately judge the condition and formulate treatment plans.

[0056] In terms of disease diagnosis, machine learning technology can discover potential disease risks and early warning signs by comprehensively analyzing and mining the patient's physiological indicators, genetic information, medical history and other data. This diagnostic method based on big data and machine learning not only improves the accuracy and reliability of diagnosis, but also provides strong support for early prevention and treatment of diseases.

[0057] In terms of treatment effect evaluation, machine learning technology can objectively evaluate the effect differences of different treatment plans by quantitatively analyzing and comparing patients' treatment effect data. This helps doctors choose treatment plans more scientifically and improve treatment effects and patient satisfaction.

[0058] However, the algorithm models of existing machine learning technologies are relatively fixed, and many medical applications directly apply existing mainstream machine learning algorithms and parameters. This fixed approach limits the flexibility and adaptability of machine learning algorithms. When faced with complex and changing medical data and scenarios, the algorithm model is prone to low fitting accuracy, low adaptability, low classification accuracy, and low fit with actual application scenarios, which in turn leads to low diagnostic accuracy.

[0059] After research, the inventor proposed a method, device, equipment and medium for constructing a diagnostic model, which includes: obtaining health status data and acupoint impedance data, the health status data and acupoint impedance data correspond one to one, and the acupoint impedance data includes acupoint information and impedance data value; according to the health status data, the acupoint impedance data is divided into healthy acupoint impedance data and unhealthy acupoint impedance data; if the healthy acupoint impedance data and the unhealthy acupoint impedance data correspond to the same acupoint information, and the significant difference between the healthy acupoint impedance data and the unhealthy acupoint impedance data is greater than the target level value, then the machine learning model is trained by the healthy acupoint impedance data and the unhealthy acupoint impedance data to obtain a diagnostic model. In this way, a diagnostic model with better algorithm model fitting accuracy, fitness, accuracy and quantification effect can be obtained, thereby improving the accuracy of diagnosis.

[0060] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0061] See also Figure 1 , which is a flow chart of a method for constructing a diagnostic model provided in an embodiment of the present application. The diagnostic method comprises:

[0062] S101: Acquire original data, where the original data at least includes acupoint impedance data and health status data.

[0063] Acupoint impedance data refers to data obtained by measuring the impedance value of acupoints on the human body. In traditional Chinese medicine theory, acupoints are channels for the circulation of qi and blood in the human body, and each acupoint is associated with a specific internal organ. By measuring acupoint impedance, acupoint impedance data can be obtained, which can indirectly reflect the physiological state and functional changes inside the human body.

[0064] Health condition data refers to the record or measurement of the user's overall health condition. In some examples, for chronic obstructive pulmonary disease research, health condition data may include "health" and "chronic obstructive pulmonary disease (COPD)". In other examples, for chronic obstructive pulmonary disease research, health condition data may also be respiratory function test data (such as blood oxygen saturation, lung function assessment, etc.), symptom score data (such as cough, wheezing, dyspnea, etc.) and / or related medical records. Through the above-mentioned respiratory function test data, symptom score data and / or related medical records, it can also be known whether the user is in a "healthy" condition or a "COPD" condition.

[0065] In some specific implementations, the original data may also include user information, such as the user's name, the user's medical number information, etc. By matching the user information with the acupoint impedance data, an acupoint impedance data table for each user may be obtained. By matching the user information with the health status data, a health status data table for each user may be obtained. See Table 1, which is a acupoint impedance data table provided in an embodiment of the present application.

[0066] Table 1

[0067] User Information Acupoint impedance data A Zhongfu point - 200Ω A Zhongfu point - 188Ω A Quchi point - 280Ω A Quchi point - 250Ω A Lieque point - 280Ω A Tianshu point - 280Ω

[0068] It should be noted that acupoint impedance data can be obtained through measuring devices such as meridian meters, and health status data can be obtained through questionnaires, clinical examinations, laboratory tests, etc. When collecting acupoint impedance data and health status data, it is necessary to ensure the authenticity and integrity of the acupoint impedance data and health status data, and comply with relevant privacy protection regulations.

[0069] S102: Cleaning the acupoint impedance data to obtain cleaned acupoint impedance data.

[0070] In some specific implementations, after obtaining the user's acupoint impedance data table, the acupoint impedance data of each acupoint of the user can be averaged to eliminate measurement errors or individual differences and obtain more representative average acupoint impedance data (i.e., acupoint impedance data after cleaning).

[0071] Subsequently, the cleaned acupoint impedance data of each acupoint of all users are integrated into a complete data set, and combined with the above health status data table to obtain a processed data table, which is easier to perform subsequent data analysis, visualization or modeling. See Table 2, which is a processed data table provided in an embodiment of the present application.

[0072] Table 2

[0073]

[0074] In other specific implementations, in addition to averaging the above-mentioned acupoint impedance data, the acupoint impedance data can also be cleaned for integrity. Integrity cleaning mainly focuses on whether there are missing values ​​or abnormal values ​​in the acupoint impedance data. The presence of missing values ​​in the acupoint impedance data means that the impedance data of some acupoints have not been correctly recorded or measured, while the presence of abnormal values ​​in the acupoint impedance data may be due to abnormal measurement results caused by equipment failure, operational errors or other reasons.

[0075] In the case where there are missing values ​​in the acupoint impedance data, the acupoint impedance data containing the missing values ​​can be deleted, or the missing values ​​can be filled; in the case where there are outliers in the acupoint impedance data, the outliers can be deleted or corrected to prevent negative impact in the subsequent model training process.

[0076] S103: Visualizing the acupoint impedance data after cleaning by using a wire box diagram method.

[0077] See also Figure 2 , which is a schematic diagram of a line box diagram provided in an embodiment of the present application. Drawing a box plot can show the distribution of the cleaned acupoint impedance data corresponding to each acupoint of multiple users in the two categories of "health" and "COPD". Dividing individuals into the two categories of "health" and "COPD" helps researchers compare and analyze the differences in the acupoint impedance data of the two groups after cleaning, helps identify the unique physiological changes or potential risk factors of COPD patients, and then provides more targeted suggestions for the prevention, diagnosis and treatment of the disease.

[0078] S104: Selecting acupoints with significant differences between healthy acupoint impedance data and unhealthy acupoint impedance data as target acupoints.

[0079] In some specific implementations, a significance test can be performed on the healthy acupoint impedance data and the unhealthy acupoint impedance data. Through the significance test, we can determine whether there is a significant difference between the two sets of data, and thus determine which acupoints' acupoint impedance data are important for distinguishing between a healthy state and an unhealthy state.

[0080] Significance testing usually involves statistical hypothesis testing, such as t-test, Mann-Whitney U test, etc. The specific test method selected depends on the distribution of acupoint impedance data and the sample size. For example, if the data meets the assumptions of normal distribution and homogeneity of variance, parametric tests such as t-test can be used; if the data does not meet these assumptions, non-parametric tests such as Mann-Whitney U test can be used. This application does not limit this.

[0081] If the significance test result is greater than the target level value (usually a preset significance level, such as 0.05 or 0.01), this means that there is a significant difference between the healthy acupoint impedance data and the unhealthy acupoint impedance data. This significant difference indicates that the impedance data of the acupoint has an important indicative role in distinguishing between healthy and unhealthy states. Therefore, we can use the acupoints corresponding to these acupoint impedance data with significant differences as target acupoints for subsequent model training processes.

[0082] S105: Training a machine learning model based on the healthy acupoint impedance data and the unhealthy acupoint impedance data of the target acupoints to obtain a diagnosis model.

[0083] In some specific implementations, the Support Vector Machine (SVM) algorithm in the supervised classification algorithm can be used to train the SVM model to obtain the diagnostic model. The SVM algorithm maps the data to a high-dimensional space by introducing a kernel function, so that an optimal hyperplane can be found in the space to separate samples of different categories.

[0084] See also Figure 3 , which is a flowchart of a training SVM model provided in an embodiment of the present application.

[0085] S31: Select healthy acupoint impedance data and / or unhealthy acupoint impedance data of the target acupoint as impedance data input into the SVM model.

[0086] S32: Divide the impedance data into a training set, a validation set, and a test set.

[0087] The training set is used to train the SVM model, that is, to let the model learn how to extract features from impedance data and classify them. The validation set is used to adjust the parameters and hyperparameters of the SVM model. By continuously adjusting the parameters and observing the performance of the SVM model on the validation set, the best parameter combination is selected. The test set is used to evaluate the final performance of the SVM model.

[0088] It should be noted that stratified sampling is required to ensure that the proportion of acupoint impedance data in healthy and unhealthy states in the training set, validation set, and test set is consistent with the proportion of impedance data, which helps the SVM model to better learn the characteristics of various types of acupoint impedance data.

[0089] S33: Select the kernel function of the SVM model.

[0090] The kernel function of the SVM model determines the mapping method of the acupoint impedance data in the high-dimensional space. For the acupoint impedance data, it may be necessary to select a suitable kernel function according to the distribution of the acupoint impedance data and the complexity of the problem. Among them, the kernel function includes a linear kernel, a polynomial kernel, and a radial basis function (RBF) kernel, etc. This application does not limit this.

[0091] S34: Use the cross-validation method to evaluate the SVM model performance and select the best parameters as the parameters of the SVM model.

[0092] Exemplarily, the cross-validation method can be a k-fold cross-validation method. Specifically, the validation set in the impedance data is divided into k subsets, k-1 subsets are used for training, and the remaining subset is used for testing. Through multiple iterations, the performance of the SVM model on different subsets of the validation set can be evaluated, and the parameter combination with the best average performance is selected as the optimal parameter of the SVM model.

[0093] S35: According to the selected kernel function and optimal parameters, the training set is input into the SVM model to train the model and obtain a diagnosis model.

[0094] According to the selected kernel function and optimal parameters, the SVM model is trained using the training set. Through multiple iterations and optimization, the SVM model will learn to extract features from the acupoint impedance data and perform classification. After the training is completed, a diagnostic model that can be used for actual diagnosis is obtained.

[0095] In some specific embodiments, after obtaining the diagnostic model, the trained diagnostic model may be tested based on the test machine obtained in S32 to obtain a confusion matrix.

[0096] See also Figure 4, which is a schematic diagram of a confusion matrix provided in an embodiment of the present application. The confusion matrix is ​​usually a 2x2 table, including the following four elements: True Positive (TP) is the number of samples predicted by the model to be positive examples, which are actually positive examples (i.e., the model diagnoses them as healthy, and they are actually healthy). False Positive (FP) is the number of samples predicted by the model to be positive examples, but they are actually negative examples (i.e., the model diagnoses them as healthy, but they are actually COPD). True Negative (TN) is the number of samples predicted by the model to be negative examples, which are actually negative examples (i.e., the model diagnoses them as COPD, and they are actually COPD). False Negative (FN) is the number of samples predicted by the model to be negative examples, but they are actually positive examples (i.e., the model diagnoses them as COPD, and they are actually healthy).

[0097] Through the above true positive examples TP, false positive examples FP, true negative examples TN, and false negative examples FN, we can get the accuracy (Accuracy), precision (Precision) and recall (Recall) of the diagnostic model. Among them, the accuracy rate indicates the proportion of samples correctly classified by the model to the total number of samples. The formula for accuracy is shown in the following formula (1). Precision measures the proportion of samples predicted by the model as positive examples that are truly positive. In other words, it reflects the reliability or accuracy of the model's prediction of positive examples. The formula for precision is shown in the following formula (2). Recall is used to measure the proportion of positive examples correctly identified by the model to all actual positive examples. In other words, recall focuses on how many samples that are actually positive examples are correctly predicted as positive examples by the model. The formula for recall is shown in the following formula (3):

[0098] Accuracy = (TP+TN) / (TP+FP+FN+TN) (1)

[0099] Precision = TP / (TP+FP)(2)

[0100] Recall = TP / (TP+FN)(3)

[0101] It should be noted that when the calculated accuracy is lower than the preset accuracy threshold, it means that the performance of the diagnostic model has not yet reached the expected standard, so the diagnostic model needs to be further adjusted and optimized, so it is necessary to continue the step of selecting the target kernel function.

[0102] It should be noted that in addition to the above indicators, other evaluation indicators such as F1 score (the harmonic mean of precision and recall) and AUC-ROC curve (calculating the true positive rate and false positive rate at different points by changing the classification threshold and drawing the curve) can be used to further evaluate the performance of the diagnosis model. These indicators can provide a more comprehensive perspective and help us to have a deeper understanding of the classification ability and stability of the diagnosis model.

[0103] In summary, the embodiment of the present application provides a method for constructing a diagnostic model, which includes: obtaining health status data and acupoint impedance data, the health status data and acupoint impedance data correspond one to one, and the acupoint impedance data includes acupoint information and impedance data value; according to the health status data, the acupoint impedance data is divided into healthy acupoint impedance data and unhealthy acupoint impedance data; if the healthy acupoint impedance data and the unhealthy acupoint impedance data correspond to the same acupoint information, and the significant difference between the healthy acupoint impedance data and the unhealthy acupoint impedance data is greater than the target level value, then the machine learning model is trained by the healthy acupoint impedance data and the unhealthy acupoint impedance data to obtain a diagnostic model. Thus, a diagnostic model with better algorithm model fitting accuracy, adaptability, accuracy, and quantification effect can be obtained, thereby improving the accuracy of diagnosis.

[0104] See also Figure 5 , which is a schematic diagram of a diagnostic model construction device provided in an embodiment of the present application. The diagnostic model construction device 500 includes: a data acquisition module 501, a data classification module 502 and a model construction module 503;

[0105] The data acquisition module 501 is used to acquire health status data and acupoint impedance data, the health status data and acupoint impedance data correspond to each other, and the acupoint impedance data includes acupoint information and impedance data values;

[0106] A data classification module 502 is used to classify the acupoint impedance data into healthy acupoint impedance data and unhealthy acupoint impedance data according to the health condition data;

[0107] The model building module 503 is used to train the machine learning model with the healthy acupoint impedance data and the unhealthy acupoint impedance data to obtain a diagnostic model if the healthy acupoint impedance data and the unhealthy acupoint impedance data correspond to the same acupoint information and the significant difference between the healthy acupoint impedance data and the unhealthy acupoint impedance data is greater than the target level value.

[0108] Optionally, the model building module 503 includes: a data differentiation module, a function selection module, a parameter selection module and a data input module;

[0109] A data differentiation module, used for distinguishing healthy acupoint impedance data and unhealthy acupoint impedance data into a training set and a validation set;

[0110] Function selection module, used to select the target kernel function of the machine learning model;

[0111] The parameter selection module is used to select the target parameters of the machine learning model through the cross-validation method;

[0112] The data input module is used to input the training set into the machine learning model. The kernel function of the machine learning model is the target kernel function, and the parameters of the machine learning model are the target parameters.

[0113] Optionally, the diagnostic model construction device 500 further includes: a test set acquisition module, a matrix acquisition module and a continued selection module;

[0114] A test set acquisition module, used to acquire a test set, the test set including healthy acupoint impedance data and / or unhealthy acupoint impedance data;

[0115] A matrix acquisition module is used to input the test set into the machine learning model to obtain a confusion matrix, which includes the accuracy of the machine learning model;

[0116] A module is continued to be selected, and is used to continue the step of selecting the target kernel function of the machine learning model if the accuracy is lower than a preset accuracy threshold.

[0117] Optionally, the model training module 503 is specifically used to: train an SVM model using a support vector machine (SVM) algorithm.

[0118] Optionally, the device for acquiring the significant difference result between healthy acupoint impedance data and unhealthy acupoint impedance data comprises: a first acquisition module and a second acquisition module;

[0119] The first acquisition module is used to perform visualization operation on healthy acupoint impedance data and unhealthy acupoint impedance data by using a wire box diagram method to obtain a wire box diagram visualization result;

[0120] The second acquisition module is used to obtain the significant difference results between the healthy acupoint impedance data and the unhealthy acupoint impedance data by analyzing the box plot visualization results.

[0121] In summary, the embodiments of the present application provide a device for constructing a diagnostic model, which can obtain a diagnostic model with better algorithm model fitting accuracy, adaptability, accuracy, and quantification effect, thereby improving the accuracy of diagnosis.

[0122] Furthermore, the present application also discloses an electronic device. Figure 6, which is a structural diagram of an electronic device provided in an embodiment of the present application. It should be noted that the content in the figure cannot be regarded as any limitation on the scope of use of the present application. The electronic device 20 may specifically include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25 and a communication bus 26. Among them, the memory 22 is used to store a computer program, and the computer program is loaded and executed by the processor 21 to implement the relevant steps in the method disclosed in any of the aforementioned embodiments. In addition, the electronic device 20 in this embodiment may specifically be an electronic computer.

[0123] In this embodiment, the power supply 23 is used to provide working voltage for each hardware device on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and the external device, and the communication protocol it follows is any communication protocol that can be applied to the technical solution of the present application, and is not specifically limited here; the input and output interface 25 is used to obtain external input data or output data to the outside world, and its specific interface type can be selected according to specific application needs and is not specifically limited here.

[0124] In addition, the memory 22, as a carrier for storing resources, can be a read-only memory, a random access memory, a disk or an optical disk, etc. The resources stored thereon can include an operating system 221, a computer program 222, etc., and the storage method can be temporary storage or permanent storage.

[0125] The operating system 221 is used to manage and control the hardware devices and computer programs 222 on the electronic device 20, which can be Windows Server, Netware, Unix, Linux, etc. In addition to including computer programs that can be used to complete the method performed by the electronic device 20 disclosed in any of the aforementioned embodiments, the computer program 222 can further include computer programs that can be used to complete other specific tasks.

[0126] Furthermore, the present application also discloses a computer-readable storage medium for storing a computer program; wherein the computer program, when executed by a processor, implements the aforementioned disclosed method. The specific steps of the method can refer to the corresponding contents disclosed in the aforementioned embodiments, and will not be repeated here.

[0127] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0128] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0129] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented directly using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0130] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0131] The technical solution provided by the present application is introduced in detail above. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea. At the same time, for general technicians in this field, according to the idea of ​​the present application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. A method for constructing a diagnostic model, characterized in that: The method comprises: Acquire health condition data and acupoint impedance data, wherein the health condition data and the acupoint impedance data correspond to each other one by one, and the acupoint impedance data includes acupoint information and impedance data values; According to the health condition data, the acupoint impedance data is divided into healthy acupoint impedance data and unhealthy acupoint impedance data; If the healthy acupoint impedance data and the unhealthy acupoint impedance data correspond to the same acupoint information, and the significant difference between the healthy acupoint impedance data and the unhealthy acupoint impedance data is greater than the target level value, a machine learning model is trained by using the healthy acupoint impedance data and the unhealthy acupoint impedance data to obtain a diagnostic model.

2. The method according to claim 1, characterized in that The training of the machine learning model by using the healthy acupoint impedance data and the unhealthy acupoint impedance data comprises: The healthy acupoint impedance data and the unhealthy acupoint impedance data are divided into a training set and a validation set; Select the objective kernel function of the machine learning model; Selecting target parameters of the machine learning model by cross-validation method; The training set is input into the machine learning model, the kernel function of the machine learning model is the target kernel function, and the parameters of the machine learning model are the target parameters.

3. The method according to claim 2, characterized in that The method further comprises: Acquire a test set, wherein the test set includes the healthy acupoint impedance data and / or the unhealthy acupoint impedance data; Inputting the test set into the machine learning model to obtain a confusion matrix, wherein the confusion matrix includes the accuracy of the machine learning model; If the accuracy is lower than the preset accuracy threshold, continue to execute the step of selecting the target kernel function of the machine learning model.

4. The method according to claim 1, characterized in that: The training of the machine learning model comprises: Use the support vector machine (SVM) algorithm to train the SVM model.

5. The method according to claim 1, characterized in that The method for obtaining the significant difference result between the healthy acupoint impedance data and the unhealthy acupoint impedance data comprises: By using a wire box diagram method, a visualization operation is performed on the healthy acupoint impedance data and the unhealthy acupoint impedance data to obtain a wire box diagram visualization result; By analyzing the box plot visualization result, the significant difference result between the healthy acupoint impedance data and the unhealthy acupoint impedance data is obtained.

6. A diagnostic model construction device, characterized in that: The device comprises: a data acquisition module, a data classification module and a model building module; The data acquisition module is used to acquire health status data and acupoint impedance data, the health status data and the acupoint impedance data correspond to each other one by one, and the acupoint impedance data includes acupoint information and impedance data values; The data classification module is used to classify the acupoint impedance data into healthy acupoint impedance data and unhealthy acupoint impedance data according to the health condition data; The model building module is used to train a machine learning model through the healthy acupoint impedance data and the unhealthy acupoint impedance data to obtain a diagnostic model if the healthy acupoint impedance data and the unhealthy acupoint impedance data correspond to the same acupoint information and the significant difference result between the healthy acupoint impedance data and the unhealthy acupoint impedance data is greater than a target level value.

7. The device according to claim 6, characterized in that The model building module includes: a data distinguishing module, a function selecting module, a parameter selecting module and a data input module; The data distinguishing module is used to distinguish the healthy acupoint impedance data and the unhealthy acupoint impedance data into a training set and a validation set; The function selection module is used to select the target kernel function of the machine learning model; The parameter selection module is used to select the target parameters of the machine learning model through a cross-validation method; The data input module is used to input the training set into the machine learning model, the kernel function of the machine learning model is the target kernel function, and the parameters of the machine learning model are the target parameters.

8. The device according to claim 7, characterized in that The device also includes: a test set acquisition module, a matrix acquisition module and a continued selection module; The test set acquisition module is used to acquire a test set, wherein the test set includes the healthy acupoint impedance data and / or the unhealthy acupoint impedance data; The matrix acquisition module is used to input the test set into the machine learning model to obtain a confusion matrix, wherein the confusion matrix includes the accuracy of the machine learning model; The continuing selection module is used to continue executing the step of selecting the target kernel function of the machine learning model if the accuracy is lower than a preset accuracy threshold.

9. An electronic device, characterized in that: include: Memory and processor; The memory is used to store programs; The processor is used to execute the program to implement each step of the method according to any one of claims 1 to 5.

10. A computer storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, each step of the method according to any one of claims 1 to 5 is implemented.