Title: On the use of random projections for dimension reduction in linear regression
Authors: Martin Slawski - George Mason Univ (United States) [presenting]
Abstract: Principal Components Regression (PCR) is a traditional tool for dimension reduction in linear regression that has been both criticized and defended. One concern about PCR is that obtaining the leading principal components tends to be computationally demanding for large data sets. While Random Projections (RPs) do not possess the optimality properties of the projection onto the leading principal subspace, they are computationally appealing and hence have become increasingly popular in recent years. We present an analysis showing that the dimension reduction offered by RPs achieves a prediction error in subsequent regression close to that of PCR, at the expense of requiring a slightly large number of RPs than PCs.