Statistics colloquium talk at the University of Iceland will be given on Wednesday, March 30th. Guðmundur Einarsson, Ph.D. Student at DTU Compute, Image Analysis & Computer Graphics will give a talk titled Accelerated Sparse Discriminant Analysis.
Abstract: Classification in high-dimensional feature spaces where interpretation and dimension reduction are of great importance is common in biological and medical applications. Often these samples are costly, e.g. microarrays, so the problems often include much more features than observations, i.e. p >> n. In this talk I will go through the formulation of the sparse optimal scoring problem for doing classification which is a generalization of Linear Discriminant Analysis when p >> n. I will also demonstrate new algorithms for solving this problem, using proximal gradient and alternating direction method of multipliers. Finally I will show an R package called accSDA that I have developed for the purpose of solving these kind of problems with R.