Computational Analysis

Our group is deeply involved in the study and implementation of methods for the processing and classification of data. In particular we develop algorithms for the improvement of signal to noise ratios (Independent Component Analysis, Common Spatial Patterns) and for data dimensionality reduction (Principal Component Analysis, Wavelet Transform), in particular for classification purposes. Moreover we develop and optimize classification algorithms (ANN, SVM, LDA) and we develop metrics for the evaluation of classification performances based on confusion matrices.