A0760
Title: Unsupervised classification method based on nonparametric functional mode estimation
Authors: Yousri Slaoui - University of Poitiers (France) [presenting]
Abstract: The focus is on an unsupervised classification problem in the framework of nonparametric functional data. We first, proposed a classification method based on a recursive estimation of the mode of the distribution of a functional random variable, this estimator is based on a pseudo-density estimator. Moreover, we study the asymptotic properties of these two estimators. We then showed the performance of the proposed unsupervised classification estimator by considering a real electroencephalography dataset. Finally, we compare our estimator to a parametric approach based on a Stochastic block Model for node-weighted networks based on two emission laws.
