B1333
Title: Enhancing performances in curves' classification: Learning from high-dimensional data via a two-step approach
Authors: Fabrizio Maturo - Universita Telem.universitas Mercatorum (Italy) [presenting]
Rosanna Verde - University of Campania Luigi Vanvitelli (Italy)
Annamaria Porreca - University of Chieti-Pescara (Italy)
Abstract: Technological progress has expanded the number and quality of devices to collect extensive data, usually recorded at temporal stamps or over time. Ordinary methods of exploring this type of information also implicate unsupervised and supervised classification strategies. However, learning from high-dimensional data is a crucial and lively topic in the statistical literature for many methodological challenges. A classification strategy is offered, combining functional data analysis with unsupervised and supervised classification. Specifically, a two-step technique is suggested. The first stage is based on extracting additional knowledge from the data using unsupervised classification employing suitable metrics. The second phase applies functional supervised classification of the new patterns learned via appropriate basis representations. The experiments on ECG data, simulation study, and comparison with the classical approaches show the effectiveness of the proposed technique and exciting refinement in terms of accuracy.
