Title: Real time dimension reduction and outlier detection
Authors: Andreas Artemiou - Cardiff University (United Kingdom) [presenting]
Yuexiao Dong - Temple University (United States)
Seung Jun Shin - Korea University (Korea, South)
Abstract: The aim is to discuss how a newly proposed algorithm for real time dimension reduction can be used for outlier detection. We first introduce the idea of SVM-based sufficient dimension reduction and then discuss how this can be extended to real time sufficient dimension reduction. Finally, we demonstrate that this algorithm can be used for efficient real time outlier detection. The good performance of the algorithm is demonstrated with simulated and real data.