CN115910326A - Bronchial asthma auxiliary diagnosis method and system based on interpretable machine learning - Google Patents
Bronchial asthma auxiliary diagnosis method and system based on interpretable machine learning Download PDFInfo
- Publication number
- CN115910326A CN115910326A CN202211543189.XA CN202211543189A CN115910326A CN 115910326 A CN115910326 A CN 115910326A CN 202211543189 A CN202211543189 A CN 202211543189A CN 115910326 A CN115910326 A CN 115910326A
- Authority
- CN
- China
- Prior art keywords
- machine learning
- model
- data
- bronchial asthma
- interpretable
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Abstract
The invention relates to an interpretable machine learning-based bronchial asthma auxiliary diagnosis method and system, and belongs to the technical field of intelligent medicine. The method comprises the steps of utilizing a preset automatic machine learning framework to carry out data preprocessing, feature engineering processing and prediction model construction, and evaluating the performance of each prediction model in a test set by adopting a cross validation method; selecting a prediction model with optimal performance according to the performance index, adjusting the hyper-parameters of the model in a predefined search space, and performing cross validation and iterative training to obtain the optimal prediction model; and performing interpretation analysis on the optimal prediction model based on an interpretable technology. The invention can relieve the user from the complicated data preprocessing, characteristic engineering, model selection and parameter tuning processes, can effectively analyze and integrate large-scale medical data, and provides reliable auxiliary diagnosis for the treatment related to the bronchial asthma.
Description
Technical Field
The invention relates to an interpretable machine learning-based bronchial asthma auxiliary diagnosis method and system, and belongs to the technical field of intelligent medicine.
Background
Bronchial asthma is one of the most common chronic respiratory diseases, and common symptoms comprise cough, expectoration, shortness of breath and the like, which seriously affect the quality of life. The pathogenesis of bronchial asthma is not clear, and a plurality of challenges are still faced in identifying and treating asthma.
With the increasing computer technology and the accumulation of data related to bronchial asthma, the systematic analysis of these data using machine learning techniques has brought clinically significant knowledge to patients. However, with the increase of the types and complexity of the machine learning algorithms, clinicians need to select corresponding frameworks for training and select appropriate models from a plurality of machine learning algorithm models, and users need to perform a lot of tedious data processing, model selection and hyper-parameter optimization. In addition, the machine learning model has a complex structure and belongs to a 'black box' model, and the output result of the model is difficult to interpret.
Disclosure of Invention
The invention aims to provide an interpretable machine learning-based bronchial asthma auxiliary diagnosis method and system, and an interpretable and reliable auxiliary diagnosis model is obtained through a small amount of manual intervention.
The invention provides a bronchial asthma auxiliary diagnosis method based on interpretable machine learning to solve the technical problem, which comprises the following steps:
1) Acquiring original data related to a bronchial asthma patient, including patient attribute information, symptoms and related medical examination data, and preprocessing the acquired data;
2) Constructing an automatic characteristic engineering frame, and inputting the preprocessed data into the automatic characteristic engineering frame for characteristic engineering processing;
3) Constructing an automatic machine learning framework and inputting the processed characteristics, wherein the machine learning framework automatically traverses a preset machine learning model and evaluates the performance index of the model;
4) Selecting a machine learning model with optimal performance according to the performance index, adjusting the hyper-parameters of the model in a predefined search space, and performing cross validation and iterative training to obtain the optimal machine learning model;
5) And (3) explaining the obtained optimal machine learning model by applying an interpretable method, and performing auxiliary diagnosis by using the optimal machine learning model.
The method comprises the steps of constructing different prediction models by utilizing a preset automatic machine learning framework, and evaluating the performance of each prediction model in a test set by adopting a cross validation method; and selecting a prediction model with optimal performance according to the performance index, adjusting the hyper-parameters of the model in a predefined search space, and performing cross validation and iterative training to obtain the optimal prediction model. The invention can relieve the user from the complicated data preprocessing, characteristic engineering, model selection and parameter tuning processes, can effectively analyze and integrate large-scale medical data, and provides reliable auxiliary diagnosis for the treatment related to the bronchial asthma.
Further, the performance index in step 3) includes: accuracy, F1-score and MCC,
wherein, accuracy represents the ratio of the samples correctly predicted and classified by the model to the total samples, and is the success rate of the model; f1-score is a harmonic mean value of combined consideration of Recall and Precision; MCC (Matthews correlation coefficient) is a comprehensive index of model performance, and the index considers true positive, true negative, false positive and false negative and is a relatively balanced index.
According to the invention, accuracy, F1-score and MCC are used as performance indexes of the evaluation model, so that the model can be accurately and comprehensively evaluated.
Further, the machine learning model preset in step 3) includes LightGBM, XGBoost, catBoost, gradient Boosting Classifier, random Forest Classifier, adaBoost, logistic Regression, extra Trees Classifier, precision Trees Classifier, ridge Classifier, linear characterization Analysis, dummy Classifier, KNN, SVM-Linear Kernel, and Naive Bayes.
The preset machine learning algorithm in the machine learning framework comprises a plurality of different algorithm models, and a reliable data source is provided for selection of subsequent models.
Further, in the step 4), an optimal prediction model is selected based on Accuracy.
According to the model selection method, model selection is carried out according to the Accuracy, and the model with the highest Accuracy is selected as the optimal prediction model according to the verification result.
Further, the method also comprises a step of verifying the optimal prediction model by using the test set.
The invention tests and verifies the obtained optimal prediction model to ensure the performance of the selected optimal prediction model.
Further, the attribute information of the patient obtained in step 1) includes age and medical history, and the related medical examination data includes laboratory examination data, lung function examination data, lung auscultation sound data, imaging data and symptom pattern data.
Further, the data preprocessing comprises repeated value deletion, abnormal value processing and missing value filling, wherein the missing value filling adopts a multiple interpolation technology based on random forests.
Because the obtained original data has the conditions of repeated values, missing values and the like, the accuracy and the generalization of the constructed model are ensured by the preprocessing mode.
Further, the feature engineering of the step 2) comprises feature preprocessing and feature selection, wherein the feature preprocessing is used for carrying out data set division, data conversion, discrete feature coding, feature normalization and unbalanced data processing on preprocessed data; feature selection includes using Person's correlation coefficient/mutual information/distance correlation, evaluating the degree of correlation between individual features and results, removing features that are less correlated with results, removing zero variance or near zero variance features, and removing co-linear features.
The invention ensures the unification of various data types by carrying out characteristic engineering processing on the preprocessed data, and is beneficial to the input of subsequent models; through feature selection and feature extraction, some redundant features are deleted, the interference of the features is reduced, and meanwhile, the precision of model training is enhanced.
The invention uses interpretability technology to interpret the obtained optimal machine learning model, and the interpretability method of the step 5) comprises the following steps: rank-based feature importance, SHAP, partial dependency graph (PDP), and Morris Sensitivity Analysis (MSA).
The invention also provides an interpretable machine learning-based bronchial asthma auxiliary diagnosis system, which comprises a processor and a memory; a memory for storing executable instructions of the processor; wherein the processor is configured to implement the interpretable machine learning-based bronchial asthma assisted diagnosis method of the present invention via execution of the executable instructions.
The method comprises the steps of utilizing a preset automatic machine learning framework to carry out data preprocessing, feature engineering processing and prediction model construction, and evaluating the performance of each prediction model in a test set by adopting a cross validation method; selecting a prediction model with optimal performance according to the performance index, adjusting the hyper-parameters of the model in a predefined search space, and performing cross validation and iterative training to obtain an optimal prediction model; and performing interpretation analysis on the optimal prediction model based on an interpretable technology. The invention can relieve the user from the complicated data preprocessing, characteristic engineering, model selection and parameter tuning processes, can effectively analyze and integrate large-scale medical data, and provides reliable auxiliary diagnosis for the treatment related to the bronchial asthma.
Drawings
FIG. 1 is a flow chart of a method for the aided diagnosis of bronchial asthma in accordance with the present invention;
FIG. 2 is a schematic diagram of the operation of the system for the auxiliary diagnosis of bronchial asthma according to the present invention;
FIG. 3 is a diagram illustrating a KS curve of an optimal predictive model obtained in an embodiment of the present invention;
FIG. 4 is a graph illustrating the significance of rank-based features of an optimal predictive model obtained in an embodiment of the invention;
FIG. 5 is a SHAP diagram of the resulting optimal predictive model in an embodiment of the invention;
FIG. 6 is a partial dependency graph of the resulting optimal prediction model in an embodiment of the present invention;
FIG. 7 is a MSA schematic of the resulting optimal prediction model in an embodiment of the invention.
Detailed Description
The following further describes embodiments of the present invention with reference to the drawings.
Embodiment of the method for auxiliary diagnosis of bronchial asthma
The method comprises the steps of firstly, acquiring original data related to a bronchial asthma patient, including patient attribute information, symptoms and related medical examination data, and preprocessing the acquired data; then, performing feature selection on the preprocessed data, and constructing a training data set based on the selected features; different prediction models are built by utilizing a preset machine learning algorithm, and cross validation is carried out on each prediction model to determine the performance index of each prediction model; selecting a machine learning model (prediction model) with optimal performance according to the performance index, adjusting the hyper-parameters of the model in a predefined search space, and performing cross validation and iterative training to obtain an optimal machine learning model (optimal prediction model); performing auxiliary diagnosis by using the obtained optimal prediction model; and performing interpretation analysis on the optimal prediction model based on an interpretable technology. The implementation flow of the method is shown in fig. 1, and the following describes the flow in detail.
1. Raw data relating to patients with bronchial asthma are acquired and preprocessed.
The data of the invention related to the acquisition of the bronchial asthma of the patient comprises the demographics, medical history, clinical symptoms, laboratory examination, pulmonary function examination, lung auscultation sound, imaging, symptom modes and the like, wherein the demographics mainly refer to the age, the sex and the like of the patient; laboratory tests including eosinophil count, lymphocyte ratio, mean corpuscular hemoglobin concentration, monocyte ratio, platelet count, erythrocyte count, leukocyte count, etc.; the imaging examination comprises chest X-ray or CT examination; symptoms include: fever, cough, dyspnea, chest distress, etc.; lung auscultation sound examination: whether wheezing sounds exist; the medical history includes diabetes history, liver disease history, kidney disease history, hypertension history and the like. Because the acquired original data may have missing values, repeated values and the like, the acquired data is preprocessed to ensure the integrity and the accuracy of the data, wherein the preprocessing mode comprises repeated value deletion, abnormal value processing, missing value filling and the like, and the missing value filling is mainly completed through a multiple interpolation technology based on a random forest algorithm.
2. And constructing an automatic characteristic engineering framework, and inputting the preprocessed data into the automatic characteristic engineering framework for characteristic engineering processing.
Because the acquired data contains more types, the data set may have high correlation and collinearity characteristics, zero variance or near-zero variance characteristics, and redundant characteristics are deleted through characteristic engineering. Therefore, the method needs to perform characteristic engineering processing on the acquired data, and the characteristic engineering processing mainly comprises characteristic preprocessing and characteristic selection.
The feature preprocessing mainly performs data set division, data conversion, discrete feature coding, feature normalization, unbalanced data processing and the like on the acquired data.
The feature selection mainly further selects feature data after the feature preprocessing, wherein the feature selection method of the invention comprises the following steps: evaluating the degree of correlation between the single characteristic and the result by utilizing the Person correlation coefficient/mutual information/distance correlation degree, and removing data which is not related to the result; deleting zero variance or near zero variance features; the collinearity features are deleted.
The feature data obtained by the processing of the step and the corresponding label data are made into a data set, and the data set is divided into a training set and a testing set according to a certain proportion.
3. And constructing an automatic machine learning framework, inputting the processed characteristics, automatically traversing preset machine learning models by using the machine learning framework, and performing cross validation on each machine learning model (also called a prediction model) to determine the performance index of each prediction model.
The preset machine learning algorithm adopted by the invention comprises LightGBM, XGboost, catBOost, gradient Boosting Classifier, random Forest Classifier, adaBoost, logistic Reguration, extra Trees Classifier, decision Tree Classifier, ridge Classifier, linear cognitive Analysis, dummy Classifier, KNN, SVM-Linear Kernel, and Naive Bayes. And constructing different prediction models by using the machine learning algorithm, training the prediction models by using a training set, performing cross validation on the obtained prediction models in a validation set by using the default parameters of the algorithm to obtain the performance indexes of the models, and sequencing based on the obtained performance indexes.
The performance indexes adopted by the invention adopt Accuracy, F1-score and MCC, wherein Accuracy represents the proportion of samples of correct prediction classification of the model in the total samples and represents the success rate of the model. The closer the Accuracy is to 1, the more accurate the model is; f1-score represents a blending average value of Recall and Precision, namely a weighted blending average value of Precision and Recall; MCC (Matthews correlation coefficient) is a comprehensive index of model performance, and the index considers true positive, true negative, false positive and false negative and is a relatively balanced index.
The performance indexes of the prediction models are obtained by verifying the prediction models by using a verification set and are shown in table 1, and the Catboost model is optimal in all models as can be seen from table 1, wherein Accuracy, F1-score and MCC are respectively 0.7487, 0.7885 and 0.4830.
TABLE 1 Performance index of each model on test set
4. And selecting a prediction model with optimal performance according to the performance index, adjusting the hyper-parameters of the model in a predefined search space, and performing cross validation and iterative training to obtain the optimal prediction model.
According to the method, the Accuracy is taken as a reference, the model with the highest Accuracy is selected as an optimal model, the optimal model automatically adjusts the hyper-parameters of the model in a predefined search space, and an iterative training is performed by adopting a cross validation method to obtain the optimal prediction model.
For the embodiment, the model with the highest Accuracy is a castboost model, and the hyper-parameters of the model after automatic tuning are as follows:
CatBoostClassifier(nan_mode＝Min,eval_metric＝Logloss,iterations＝270,sampling_f requency＝PerTree,leaf_estimation_method＝Newton,grow_policy＝SymmetricTree,penalties_coefficient＝1,boosting_type＝Plain,
model_shrink_mode＝Constant,feature_border_type＝GreedyLogSum,bayesian_matrix_reg＝0.1,force_unit_auto_pair_weights＝False,l2_leaf_reg＝8,random_strength＝0.1,rsm＝1,boost_from_average＝False,model_size_reg＝0.5,pool_metainfo_options＝{'tags':{}},subsample＝0.8,use_best_model＝False,class_names＝[1,2],random_seed＝5252,depth＝6,posterior_sampling＝False,border_count＝254,classes_count＝0,auto_class_weights＝None,sparse_features_conflict_fraction＝0,leaf_estimation_backtracking＝AnyImpro vement,best_model_min_trees＝1,model_shrink_rate＝0,min_data_in_leaf＝1,loss_fu nction＝Logloss,learning_rate＝0.001,score_function＝Cosine,task_type＝CPU,leaf_es timation_iterations＝10,bootstrap_type＝MVS,max_leaves＝64)。
5. and (5) performing super-parameter tuning to obtain the performance index of the optimal prediction model.
The method utilizes a test set to verify a model for adjusting the hyper-parameters to obtain performance indexes of the prediction model, such as Accuracy, F1-score, MCC and the like, which are 0.7652, 0.7879 and 0.5349 respectively; the Accuracy of the prediction model after the super-parameter adjustment is improved by 2.2 percent compared with the original model.
In order to better illustrate the performance of the model, the performance of the optimal prediction model is visually analyzed, and the automatic machine learning framework can automatically draw performance indexes of the model, including an ROC curve, a confusion matrix, precision-Recall, variable importance, a KS curve and the like. The KS curve is also called Lorentz curve and is drawn according to TPR (true positive rate) and FPR (false positive rate), when in drawing, the threshold value is taken as a horizontal axis, the TPR and the FPR are taken as vertical axes, two curves are obtained, the threshold value corresponding to the position with the farthest distance between the two curves is the threshold value which can divide the model most, and the adopted formula is as follows: KS = max (TPR-FPR), which is the maximum value of the difference between TPR and FPR. The KS value is used to evaluate the discriminative power of the model, the greater the KS value, the greater the discriminative power of the model. The value range of the KS value is [0,1], and generally, the KS is more than 0.2, so that the model has better prediction accuracy. For this example, the KS curve obtained is shown in fig. 3, and the KS value is 0.691 when the threshold value is 0.483, which has a good prediction ability.
6. And performing interpretability analysis on the optimal model subjected to the super-parameter adjustment, so that the model can be further understood and reference is provided for medical decision.
Interpretability is important in the medical field, and a medical diagnostic system must be transparent, interpretable to be able to gain trust in doctors and patients. Interpretable techniques include: rank-based feature importance, SHAP, partial dependency graph (PDP), and Morris Sensitivity Analysis (MSA).
The feature importance of the optimal prediction model obtained in the embodiment based on arrangement in the data set is shown in fig. 4, and it can be seen that LBXBCD, ENQ090_1 and WTMEC5YR have the greatest influence on the performance of the constructed model. The distribution of the SHAP of the features of the optimal prediction model in the data set obtained in this embodiment is shown in FIG. 5, where the vertical axis is the ranking of the features according to the sum of SHAP values of all samples, and the horizontal axis is the SHAP value (the distribution of the influence of the features on the model output); each point represents a sample, the sample size is piled up longitudinally, and the color represents a characteristic value (dark color corresponds to high value, light color corresponds to low value); it can be seen that the features RQD070_1, ENQ090_1, SPXNFCV, and SPXNFEV1 contribute significantly to the model.
The dependency between features and model objectives can be shown by a dependency graph, as shown in fig. 6, which is the influence of the RIDAGEYR features in the optimal prediction model obtained in this embodiment on model prediction, from which it can be seen that RIDAGEYR is almost independent of bronchial asthma. The global sensitivity of the optimal prediction model obtained by the invention is evaluated by adopting an MSA method, and the result is shown in FIG. 7, wherein RDQ070_1, ENQ090_1 and SPXNFVC have larger influence on the model and are basically consistent with the SHAP analysis result.
Through the process, the method can be effectively used for analyzing and integrating large-scale medical data, so that a doctor can make a decision and guide clinical practice, accurate medical treatment is brought to patients with bronchial asthma, and the method has important clinical significance.
Embodiment of auxiliary diagnosis system for bronchial asthma
The system for auxiliary diagnosis of bronchial asthma of the present invention comprises a processor and a memory, the processor executes a computer program stored by the memory to implement the method of the present invention implementing the above-described method embodiment, the principle of which is shown in fig. 2. That is, the method in the above method embodiment should be understood that the flow of the method for the auxiliary diagnosis of bronchial asthma may be implemented by computer program instructions. These computer program instructions may be provided to a processor such that execution of the instructions by the processor results in the implementation of the functions specified in the method flow described above.
The processor referred to in this embodiment refers to a processing device such as a microprocessor MCU or a programmable logic device FPGA; the memory referred to in this embodiment includes a physical device for storing information, and generally, information is digitized and then stored in a medium using an electric, magnetic, optical, or the like. For example: various memories for storing information by using an electric energy mode, such as RAM, ROM and the like; various memories for storing information by magnetic energy, such as hard disk, floppy disk, magnetic tape, magnetic core memory, bubble memory, and U disk; various types of memory, CD or DVD, that store information optically. Of course, there are other ways of memory, such as quantum memory, graphene memory, and so forth.
The device composed of the memory, the processor and the computer program is realized by the processor executing corresponding program instructions in the computer, and the processor can be loaded with various operating systems, such as a Windows operating system, a Linux system, an Android system, an iOS system and the like.
As other embodiments, the device can also comprise a display, and the display is used for displaying the selection result for the staff to refer to.
The foregoing is merely a preferred embodiment of the invention and is not intended to limit the invention, as many variations and modifications are possible without departing from the scope of the invention. It is therefore intended that the foregoing detailed description be regarded as illustrative rather than limiting, and that it be understood that it is the following claims, including all equivalents, that are intended to define the spirit and scope of this invention. The above examples are to be construed as merely illustrative and not limitative of the remainder of the disclosure. After reading the description of the invention, the skilled person can make various changes or modifications to the invention, and these equivalent changes and modifications also fall into the scope of the invention defined by the claims.
Claims (10)
1. An auxiliary diagnosis method for bronchial asthma based on interpretable machine learning, which is characterized by comprising the following steps:
1) Acquiring original data related to a bronchial asthma patient, including patient attribute information, symptoms and related medical examination data, and preprocessing the acquired data;
2) Constructing an automatic characteristic engineering frame, and inputting the preprocessed data into the automatic characteristic engineering frame for characteristic engineering processing;
3) Constructing an automatic machine learning framework and inputting the processed characteristics, wherein the machine learning framework automatically traverses a preset machine learning model and evaluates the performance index of the model;
4) Selecting a machine learning model with optimal performance according to the performance indexes, adjusting the hyper-parameters of the model in a predefined search space, and performing cross validation and iterative training to obtain the optimal machine learning model;
5) And (4) explaining the obtained optimal machine learning model by using an interpretability method, and performing auxiliary diagnosis by using the optimal machine learning model.
2. The method for auxiliary diagnosis of bronchial asthma based on interpretable machine learning of claim 1, wherein the main performance indicators in step 3) include: accuracy, F1-score and MCC, wherein the Accuracy represents the ratio of samples of correct prediction classification of the model to total samples, and is the success rate of the model; f1-score is a harmonic mean value of combined consideration of Recall and Precision; MCC (Matthews correlation coefficient) is a comprehensive index of model performance, and the index considers true positive, true negative, false positive and false negative and is a relatively balanced index.
3. The method for auxiliary diagnosis of bronchial asthma based on interpretable machine learning of claim 1, wherein the machine learning model preset in step 3) comprises LightGBM, XGBoost, catBoost, gradient Boosting Classifier, random Forest Classifier, adaBoost, logistic Regression, extra Trees Classifier, decision Tree Classifier, ridge Classifier, linear Discriminant Analysis, dummy Classifier, KNN, SVM-Linear Kernel, and Naive Bayes.
4. The method for auxiliary diagnosis of bronchial asthma based on interpretable machine learning of claim 2, wherein in the step 4), an optimal prediction model is selected based on Accuracy.
5. The method for aided diagnosis of bronchial asthma based on interpretable machine learning of claim 4, further comprising the step of verifying the best predictive model using a test set.
6. The method for assisting diagnosis of bronchial asthma based on interpretable machine learning of claim 1, wherein the patient attribute information obtained in step 1) includes age and medical history, and the related medical examination data includes laboratory examination data, lung function examination data, lung auscultation sound data, imaging data and symptom pattern data.
7. The method as claimed in claim 1, wherein the data preprocessing comprises duplicate value deletion, outlier processing and missing value filling, wherein the missing value filling employs multiple interpolation techniques based on random forests.
8. The method for assisting diagnosis of bronchial asthma based on interpretable machine learning according to claim 1, wherein the feature engineering of step 2) comprises feature preprocessing and feature selection, wherein the feature preprocessing is used for carrying out data set partitioning, data transformation, discrete feature coding, feature normalization and unbalanced data processing on preprocessed data; feature selection includes using Person correlation coefficient/mutual information/distance correlation, evaluating the degree of correlation between individual features and results, removing features that are less correlated with results, removing zero variance or near zero variance features, and removing collinearity features.
9. The method for auxiliary diagnosis of bronchial asthma based on interpretable machine learning of claim 1, wherein the interpretable method of step 5) comprises: rank-based feature importance, SHAP, partial dependency graph (PDP), and Morris Sensitivity Analysis (MSA).
10. An interpretable machine learning-based bronchial asthma aided diagnosis system, comprising a processor and a memory; a memory for storing executable instructions of the processor; wherein the processor is configured to implement the interpretable machine learning-based bronchial asthma assisted diagnostic method of any one of claims 1-9 via execution of the executable instructions.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211543189.XA CN115910326A (en) | 2022-12-02 | 2022-12-02 | Bronchial asthma auxiliary diagnosis method and system based on interpretable machine learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211543189.XA CN115910326A (en) | 2022-12-02 | 2022-12-02 | Bronchial asthma auxiliary diagnosis method and system based on interpretable machine learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115910326A true CN115910326A (en) | 2023-04-04 |
Family
ID=86478581
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211543189.XA Pending CN115910326A (en) | 2022-12-02 | 2022-12-02 | Bronchial asthma auxiliary diagnosis method and system based on interpretable machine learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115910326A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111938652A (en) * | 2020-08-24 | 2020-11-17 | 上海市第一人民医院 | Application of artificial intelligence diagnosis mode in early diagnosis of bronchial asthma |
-
2022
- 2022-12-02 CN CN202211543189.XA patent/CN115910326A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111938652A (en) * | 2020-08-24 | 2020-11-17 | 上海市第一人民医院 | Application of artificial intelligence diagnosis mode in early diagnosis of bronchial asthma |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Azar et al. | Decision tree classifiers for automated medical diagnosis | |
| CN110051324B (en) | Method and system for predicting death rate of acute respiratory distress syndrome | |
| US20220093215A1 (en) | Discovering genomes to use in machine learning techniques | |
| CN111598867B (en) | Method, apparatus, and computer-readable storage medium for detecting specific facial syndrome | |
| CN115131642B (en) | Multi-modal medical data fusion system based on multi-view subspace clustering | |
| CN114220540A (en) | Construction method and application of diabetic nephropathy risk prediction model | |
| JP2023532292A (en) | Machine learning based medical data checker | |
| CN115910326A (en) | Bronchial asthma auxiliary diagnosis method and system based on interpretable machine learning | |
| CN112183572A (en) | Method and device for generating prediction model for predicting pneumonia severity | |
| Shrestha et al. | Supervised machine learning for early predicting the sepsis patient: modified mean imputation and modified chi-square feature selection | |
| CN111986814A (en) | Modeling method of lupus nephritis prediction model of lupus erythematosus patient | |
| CN113408603B (en) | Coronary artery stenosis degree identification method based on multi-classifier fusion | |
| CN112508884A (en) | Comprehensive detection device and method for cancerous region | |
| CN117195027A (en) | Cluster weighted clustering integration method based on member selection | |
| CN111080625B (en) | Training method and training device for lung image strip and rope detection model | |
| Kotiyal et al. | Diabetic Retinopathy Binary Image Classification Using Pyspark | |
| CN115600091B (en) | Classification model recommendation method and device based on multi-modal feature fusion | |
| CN116226629B (en) | Multi-model feature selection method and system based on feature contribution | |
| Setiawan et al. | ANALYSIS OF CLASSIFICATION OF LUNG CANCER USING THE DECISION TREE CLASSIFIER METHOD | |
| Miranda | HYTEA-HYBRID TREE EVOLUTIONARY ALGORITHM FOR HEARING LOSS DIAGNOSIS | |
| CN112530591B (en) | Method for generating auscultation test vocabulary and storage equipment | |
| Anil Kumar | Diabetic Retinopathy stages Classifications using Lesion features and CNN Models | |
| Reddy | Classifying Blood Glucose Levels Through Noninvasive Features | |
| CN117253073A (en) | Method and device for judging efficient classification of medical images and storage medium | |
| Manickavasagam et al. | Deep Belief Network for Lung Nodule Segmentation and Cancer Detection. |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |