Functional data clustering using principal curve methods

In this paper we propose a novel clustering method for functional data based on the principal curve approach. By are approximated using component analysis (FPCA) and is then performed scores. The proposed makes use of nonparametric curves to summarize features scores extracted from original data, probabilistic model combined with Bayesian Information Criterion employed automatically simultaneously find appropriate number features, optimal degree smoothing corresponding cluster members. simulation studies show that outperforms existing approaches considered. capability also demonstrated by applications in human mortality fertility data.