Sparse Boosted PLS

摘要：

L2Boosting and PLS are effective methods for regression in high dimensional case, where a sparse model is usually preferred. In fact, the sparsity of the components and the number of the iteration play the key roles to construct sparse model of good prediction performance. In this paper, combining L2Boosting and sparse PLS, we propose a new method called sparse boosted PLS (SBPLS). This method uses the sparse components to participate the L2Booting iteration. Based on corrected AIC, we select the degree of sparsity of the components and the number of iteration. Simulations show that this method has good prediction and can make variable selection.

关键词： partial least square L2Boosting sparse partial least square

作者: Zhao Xiuli Zhao Junlong Wu Xizhi

作者单位: Statistic School, Renmin University of China, Beijing, 100872 Department of mathematics, Beihang University, LMIB of the Ministry of Education of China: Beijing,

会议类型: 国际会议

会议名称: The 6th International Conference on Partial Least Squares and Related Methods(第六届偏最小二乘及相关方法国际会议)

会议地点: 北京

会议语种:英文

页码: 132-136

在线出版日期: 2009-09-04（万方平台首次上网日期，不代表论文的发表时间）