A Novel Ensemble Learning Method for Relevance Vector Regression

摘要：

Relevance vector machine(RVM) is a state-of-the-art technique for regression and classification, as a sparse Bayesian extension version of the support vector machine. To improve the prediction performance of RVM, ensemble learning is introduced. The paper proposes a novel relevance vector regression(RVR) ensemble learning method. Firstly, bootstrap sampling used by Bagging is employed to manipulate the training set. Secondly, RVR learners are trained with different multiple kernel function. Finally, taking the predicting accuracy of selected ensemble as the optimization object, memetic algorithm is applied to obtain the final selective ensemble. Experiment result shows the proposed method can improve the generalization performance compared with single RVR, Bagging RVR, Boosting RVR.

关键词： relevance vector regression bagging multiple kernel selection selective ensemble memetic algorithm

作者: Wu Bing Liang Jia-hong Chen Ling Hu Zhi-wei

作者单位: College of Mechanical Engineering and Automation, National University of Defense Technology Changsha College of Mechanical Engineering and Automation, National niversity of Defense Technology Changsha,

会议类型: 国际会议

会议名称: 2010 Third Pacific-Asia Conference on Web Mining and Web-based Application(2010年第三届web挖掘和基于web应用亚太会议 WMWA 2010)

会议地点: 桂林

会议语种:英文

页码: 362-365

在线出版日期: 2010-11-17（万方平台首次上网日期，不代表论文的发表时间）