Prediction of osteoporotic fracture based on ensemble learning(PDF)
《中国医学物理学杂志》[ISSN:1005-202X/CN:44-1351/R]
- Issue:
- 2021年第2期
- Page:
- 254-258
- Research Field:
- 医学人工智能
- Publishing date:
Info
- Title:
- Prediction of osteoporotic fracture based on ensemble learning
- Author(s):
- CHEN Wanqi; LIN Yong
- School of Medical Instrument and Food Engineering, University of Shanghai for Science and Technology, Shanghai 200093, China
- Keywords:
- Keywords: osteoporotic fracture machine learning ensemble learning classification prediction ten-fold cross validation
- PACS:
- R318；R683
- DOI:
- DOI:10.3969/j.issn.1005-202X.2021.02.023
- Abstract:
- Abstract: Osteoporotic fracture is one of the important causes of morbidity and death in the elderly. It is necessary to establish an efficient predictive model to provide diagnosis and treatment suggestions for the elderly as soon as possible. In the experiment, Stacking is used to construct a heterogeneous classifier EtDtb-S which uses 16 highly-correlated features as feature vectors, and selects extreme random trees and decision tree-based bagging ensemble models as primary learners, and logistic regression as the secondary learner for ensemble learning. Experimental verification compares EtDtb-S with single model and isomorphic classifiers for osteoporotic fracture prediction. The results show that the prediction accuracy of the heterogeneous classifier is increased by 2.8% and 1.5% as compared with the optimal single model and the optimal isomorphic classifier, respectively. The proposed method has better prediction of osteoporotic fracture.
Last Update: 2021-02-04