Title: A nonparametric functional clustering of mouse ultrasonic vocalization data
Authors: Xiaoling Dou - Waseda University (Japan) [presenting]
Abstract: A nonparametric method of functional clustering is proposed. The method is based on functional principle component analysis. We calculate the principle component score for each functional data, and estimate the density of the principle component score. Then we find the mode for each principle component, and classify the functional data by calculating the distance of the functional data from the mode. A real dataset of mouse ultrasonic vocalization is illustrated.