青岛大学学报（自然科学版）2023，Vol.36Issue(4)：35-40,6.DOI:10.3969/j.issn.1006-1037.2023.04.06
基于布谷鸟搜索的XGBoost算法优化及应用研究
The Optimization and Application on XGBoost Algorithm Based on Cuckoo Search
摘要
Abstract
In order to improve the prediction accuracy of XGBoost algorithm,the cuckoo search algorithm was used to globally optimize hyperparameters in XGBoost algorithm:learning rate,the minimum loss of output node splitting,the maximum depth of tree model and the number of base classifiers.The CS-XG-Boost model was built to train the dataset.The results show that the accuracy,precision,F1-score and AUC of the CS-XGBoost income classification model obtain 95.67%,97.17%,95.56%and 97.96%,which are higher than Logistic regression model,support vector machine,random forest,XGBoost and XGBoost algorithm based on grid search.The coefficient of determination,root mean square error and mean absolute error of the CS-XGBoost housing price prediction model are 0.905 5,2.943 5 and 2.165 4.Compared with XGBoost algorithm,CS-XGBoost algorithm can effectively improve the prediction accura-cy.关键词
XGBoost/布谷鸟搜索/分类预测/回归预测Key words
XGBoost/cuckoo search/classification prediction/regression prediction分类
信息技术与安全科学引用本文复制引用
李欣玲,李莉莉,周楷贺..基于布谷鸟搜索的XGBoost算法优化及应用研究[J].青岛大学学报（自然科学版）,2023,36(4):35-40,6.基金项目
国家社科基金(批准号:2019BTJ028)资助 （批准号:2019BTJ028）
山东省金融应用重点研究项目(批准号:2020-JRZZ-03)资助. （批准号:2020-JRZZ-03）
