摘要：Projection Pursuit Regression (PPR) has played an important role in the development of statistics and machine learning. However, as a statistical learning method, PPR has not yet demonstrated an accuracy comparable to other methods such as Random Forests (RF) and Artificial Neural Networks (ANN). In this paper, we revisit the estimation of PPR and propose a greedy algorithm and an ensemble approach via feature bagging, hereafter referred to as ePPR. Compared to Random Forest (RF), ePPR has two main advantages: (1) its theoretical consistency can be proved for more general regression functions, as long as they are continuous, and higher consistency rates can be obtained; and (2) ePPR does not split the samples, so each term of the PPR is estimated using the whole data, which makes the estimation more efficient and guarantees the smoothness of the estimator. ePPR is also easier to tune and train than ANN. Extensive comparisons on real data sets show that ePPR is noticeably more efficient in regression and classification than RF and other competitors.

报告人简介：夏应存,新加坡国立大学统计与数据科学系教授。研究兴趣包括非参数回归，高维数据分析，疾病传播统计建模等。研究成果发表在AOS，JASA，JRSSB, Biometrika, JOE, PNAS等期刊. Nature News等多个学术媒体对其提出的疾病跨域传播模型做了专题报道。JRSSB，Statistical Science和Statistica Sinica对其论文进行了公开讨论。 夏应存曾在暨南大学工作多年，获国务院侨办颁发的“优秀教师”称号.