Abstract A robust approach for clustering functional directional data is proposed. The proposal adapts “impartial trimming” techniques to this particular framework. Impartial trimming uses the dataset itself tell us which appears be most outlying curves. feasible algorithm proposed its practical implementation justified by some theoretical properties. “warping” also introduced allows including ...

Abstract A new model-based procedure is developed for sparse clustering of functional data that aims to classify a sample curves into homogeneous groups while jointly detecting the most informative portions domain. The proposed method referred as and smooth (SaS-Funclust) relies on general Gaussian mixture model whose parameters are estimated by maximizing log-likelihood function penalized with...

Cluster analysis seeks to assign objects with similar characteristics into groups called clusters so that within a group are each other and dissimilar in groups. Spectral clustering has been shown perform well different scenarios on continuous data: it can detect convex non-convex clusters, overlapping clusters. However, the constraint data be limiting real applications where often of mixed-typ...