摘要
目的由于传统机器学习算法的分类能力较低,不足以辅助临床诊断,本研究将分类功能强大的集成学习与医疗诊断相结合,提高诊断准确率和召回率。方法研究应用集成学习的随机森林算法和Xgboost算法来提高模型准确率和召回率,并利用交叉验证和网格搜索提高模型拟合能力。结果通过对比随机森林模型、Xgboost模型和传统机器学习的决策树模型,研究得出,集成学习极大地提高了乳腺癌诊断的准确率和召回率,准确率从0.92提高至0.96,召回率从0.90提高至0.97和0.99。结论将集成学习算法与实际医疗诊断技术相结合具有实际的研究意义,可以进一步将两种领域相结合,以提高医疗诊断的效率和准确率。
Objective Due to the low classification ability of traditional machine learning algorithm,it is not enough to assist clinical diagnosis.This project combines the ensemble learning with strong classification function for medical diagnosis to improve the diagnosis accuracy and recall rate.Methods In this project,the random forest algorithm and Xgboost algorithm of ensemble learning were applied to improve the accuracy and recall rate of the model.Cross validation and grid search were used to improve the model fitting ability.Results By comparing random forest model,Xgboost model and decision tree model of traditional machine learning,we can see that ensemble learning has greatly improved the accuracy and recall rate of breast cancer diagnosis.The accuracy rate increased from 0.92 to 0.96,and the recall rate increased from 0.90 to 0.97 and 0.99.Conclusion It has practical research significance to combine the ensemble learning algorithm with the actual medical diagnosis technology.We can further combine the two fields to improve the effectiveness of medical diagnosis.
作者
邓卓
苏秉华
张凯
DENG Zhuo;SU Binghua;ZHANG Kai(Key Laboratory of Photoelectric Imaging and System,Ministry of Education,Zhuhai College of Beijing Institute of Technology,Zhuhai Guangdong 519088,China;Beijing Institute of Technology,Beijing 100081,China)
出处
《中国医疗设备》
2020年第12期59-62,共4页
China Medical Devices
基金
珠海市光电信息技术协同创新中心基金项目。
