基于Stacking集成学习的肺癌患者存活性预测模型研究  

The Prediction Model of Survival Activity in Lung Cancer Patients Based on Stacking Integrated Learning

作　　者：纪江明 赵玲 JI Jiangming;ZHAO Ling(School of Economics and Management,Huzhou University,Huzhou 313000,China;School of Information Engineering,Huzhou University,Huzhou 313000,China)

机构地区：[1]湖州师范学院经济管理学院,浙江湖州313000 [2]湖州师范学院信息工程学院,浙江湖州313000

出　　处：《湖州师范学院学报》2024年第2期83-91,共9页Journal of Huzhou University

基　　金：浙江省哲学社会科学规划项目(24NDJC024YB);国家社会科学基金项目(16BRK012);湖州科技局公益性应用研究项目(2021GZ56)。

摘　　要：为提高肺癌患者存活性预测的准确率,提出一种基于Stacking集成学习的肺癌患者存活性预测模型.先对数据集进行预处理、特征选择、变量转换等,然后以XGBoost(eXtreme Gradient Boosting)、SVM(Support Vector Machine)和LR(Logistic Regression)3种算法为基学习器,以朴素贝叶斯为元学习器构造模型,再运用Grid Search网格搜索方法优化超参数,并利用交叉验证方法对SEER公开的肺癌数据集进行仿真实验.研究结果表明,该模型的预测准确率达85%,比单一模型高10%.该模型在肺癌患者存活性预测上有着更好的准确性和解释性,可以很好地为肺癌患者预后提供决策支持,以弥补经验的不足.In order to improve the accuracy of survivability prediction for lung cancer patients,a survivability prediction model for lung cancer patients based on Stacking integrated learning is proposed.Firstly,the data set is preprocessed,feature selection,variable conversion,etc.Furthermore,XGBoost(eXtreme Gradient Boosting),SVM(Support Vector Machine)and LR(Logistic Regression)algorithms were used as the based learning algorithms and naive Bayes as the meta-learning algorithms to construct the model.Secondly,the Grid Search method was used to optimize the hyperparameters and cross validation method to conduct simulation experiments on the lung cancer data set disclosed by SEER.The results show that the prediction accuracy of this model is 85%,which is 10%higher than that of the single model.Therefore,this model has better accuracy and interpretation in the prediction of survival activity of lung cancer patients,and can well provide decision support for the prognosis of lung cancer patients to make up for the lack of experience.

关 键 词：存活性预测 肺癌患者 集成学习 交叉验证 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]