Title: A first approach to real time sufficient dimension reduction
Authors: Andreas Artemiou - Cardiff University (United Kingdom) [presenting]
Seung Jun Shin - Korea University (Korea, South)
Yuexiao Dong - Temple University (United States)
Abstract: A work in progress to create new effective and computationally efficient algorithms for real time sufficient dimension reduction (SDR) is presented. In an era where massive datasets are becoming the norm, we present a method where after an initial estimation based on our original data, when new data are collected the reduced subspace can be estimated using our proposed methods using only new data and completely ignoring the old data. We will demonstrate how this algorithm can be extended to sparse SDR.