Dimensionality reduction based on ICA for regression problems

Abstract

In manipulating data such as in supervised learning, we often extract new features from the original input variables for the purpose of reducing the dimensions of input space and achieving better performances. In this paper, we show how standard algorithms for independent component analysis (ICA) can be extended to extract attributes for regression problems. The advantage is that general ICA algorithms become available to a task of dimensionality reduction for regression problems by maximizing the joint mutual information between target variable and new attributes. We applied the proposed method to a couple of real world regression problems as well as some artificial problems and compared the performances with those of other conventional methods. Experimental results show that the proposed method can efficiently reduce the dimension of input space without degrading the regression performance.