A novel wavelet transform based on polar coordinates for datamining applications

Abstract

In this paper, we propose a novel wavelet transform based on the polar coordinates for datamining applications. In general, the Harr wavelet transform has been popularly used for data decomposition. However, the Harr wavelet transform shows the poor performance for the locally distributed data which are clustered around certain values, since it uses the averages as representatives for data decomposition. The proposed wavelet transform is based on the the polar coordinates which is not affected by the averages and is more suitable than the Harr wavelet transform for data decomposition of the locally distributed data.