YAO Shuaijun,YAN Jinglai,DU Caifeng,et al.Research on Intelligent Diagnosis Model of Traditional Chinese Medicine based on Ensemble Learning: Taking Perimenopausal Syndrome as an Example[J].Journal of Traditional Chinese Medicine,2023,64(06):572-580.
YAO Shuaijun,YAN Jinglai,DU Caifeng,et al.Research on Intelligent Diagnosis Model of Traditional Chinese Medicine based on Ensemble Learning: Taking Perimenopausal Syndrome as an Example[J].Journal of Traditional Chinese Medicine,2023,64(06):572-580.DOI: 10.13288/j.11-2166/r.2023.06.007.
Research on Intelligent Diagnosis Model of Traditional Chinese Medicine based on Ensemble Learning: Taking Perimenopausal Syndrome as an Example
To set up a traditional Chinese medicine (TCM) intelligent diagnosis model of perimenopausal syndrome based on the model fusion method in ensemble learning.
Methods
2
The databases including CNKI,WanFang Data,VIP,CBM as well as the ancient and modern medical case cloud platform were searched from inception date to May 20th, 2022,to collect the 1021 data samples of medical cases of perimenopausal syndrome treated by TCM and then establish the data set. The 1021 data samples were divided into training set(n = 714) and test set(n = 307) by a ratio of 7∶3. Random forest, support vector machine and artificial neural network were used to build the syndrome differentiation diagnosis models of perimenopausal syndrome,and the above classifier models were used to learn the data set and make predictions. Then the hyper-parameters of the models were optimized by grid search method,and the three classifier models were fused by heterogeneous integration method to form a strong learner. And the accuracy, precsion,recall, f1 score,and the area under the curve (AUC) value were taken as the evaluation indicators of models.
Results
2
The accuracy rate of random forest, support vector machine, neural network and ensemble model was 0.84, 0.86, 0.85, and 0.87, respectively; the corresponding precsion rates were 0.85, 0.86, 0.85 and 0.87; the recall rates were 0.84, 0.86, 0.85, and 0.87, respectively; the f1 scores were 0.84, 0.86, 0.85 and 0.87;the AUC values were 0.98, 0.97, 0.97 and 0.98. The ensemble model was optimal in all indicators.
Conclusion
2
The ensemble model which has better stability can synthesize the advantages of each weak supervision model, and its performance and generalization ability are better than that of a single learner, thereby having significant reference value for clinical diagnosis in TCM.
LUNDBERG S , LEE SI . A unified approach to interpreting model predictions [EB/OL].( 2017-12-04 )[ 2022-05-01 ]. https://proceedings.neurips.cc/paper/2017/file/8a20a8621978632d76c43dfd28b67767-Paper.pdf https://proceedings.neurips.cc/paper/2017/file/8a20a8621978632d76c43dfd28b67767-Paper.pdf .
LI R , SHINDE A , LIU A , et al . Machine learning-based interpretation and visualization of nonlinear interactions in prostate cancer survival [J]. JCO Clin Cancer Inform , 2020 : 637 - 646 . doi: 10.1200/CCI.20.00002 http://dx.doi.org/10.1200/CCI.20.00002 .
The Research Progress and Development Strategies of Traditional Chinese Medicine Diagnosis Empowered by Artificial Intelligence
An AI-Driven Research Pathway for Upholding Fundamental Principles and Breaking New Ground of Traditional Chinese Medicine Under the Guidance of Complexity Science
Construction and Clinical Validation of a Deep Learning-Based Automatic Measurement Model for Palmar Tilt and Radial Inclination in Distal Radius Fractures
Construction and Application of a Real-World Cohort of Community-Acquired Pneumonia Based on a Multimodal Large-Scale Traditional Chinese Medicine Big Data Platform
The Current Issues and Thoughts on the Empowerment of Famous Doctors' Experience Inheritance by Artificial Intelligence
Related Author
ZHU Wenjun
TANG Manshi
SHE Kaijie
TANG Zihao
HUANG Minyi
YUAN Naijun
MA Qingyu
CHEN Jiaxu
Related Institution
School of Traditional Chinese Medicine,Beijing University of Chinese Medicine
School of Traditional Chinese Medicine,Jinan University
School of Intelligent Science and Technology,University of Science and Technology Beijing
Tsinghua University
Guang 'anmen Hospital,China Academy of Chinese Medical Sciences