DCA algorithm for clusterwise linear regression and its comparison

Clusterwise linear regression consists of finding a number of linear regression functions each approximating a subset of the data. It is a combination of two techniques: clustering and regression. We introduce an algorithm for solving the clusterwise linear regression problem using its nonsmooth optimization formulation and difference of convex representation. The algorithm is tested using real world data sets and compared with other clusterwise linear regression algorithms

Keywords:

Nonconvex optimization , DC optimization , Nonsmooth optimization , Clusterwise linear regression