A Clustering Method for Misaligned Curves

Abstract

In this talk, I will present a clustering algorithm for misaligned curves. If curves show a common pattern after alignment using nonlinear warping functions, they are considered similar. In the proposed clustering method, curves are updated such that similar curves become more similar and it is easier to cluster the curves. I will show some simulation results to demonstrate the differences between the proposed method and k-means clustering, and explain why the differences are expected. Two real data applications will be presented. This talk is based on joint work with Tzee-Ming Huang and Su-Fen Yang at National Chengchi University.