Improved Single Target Accuracy Using Feature Selection for Multitarget Regression

Abstract

Multi- Target Regression (MTR) has recently gained a great deal of attention due to a large number of realworld applications being identified. Several MTR algorithms have been proposed and more continue to be developed. The main objective of all MTR prediction strategies is to use inter-target dependencies to improve prediction efficiency. In this research, we explore the affect of feature selection on MTR problems. We observe that using ordinary feature selection techniques with Single Target (ST) approach can give at par results compared to MTR methods, and in some cases, even surpass them.