Principal Component Analysis of high dimensional data often runs into time and memory limitations. This is especially the case if the dimension and the number of data set elements is of about the same size. We propose a new method to calculate Principal Components based on Compressive Sensing. Compressive Sensing can be interpreted as a new method for data compression with a number of positive ...

This paper presents the motivation for and the preliminary theoretical investigations of the PhD project by the first author. The objective of the research is to propose and to experimentally verify the approach of application of eigendecomposition in principal component analysis for topic modelling of short text document collections. The main hypothesis examined in this project, is that princi...

In this paper, we propose to employ an extension to the natural gradient algorithm for robust Independent Component Analysis against outliers. The standard natural gradient algorithm does not exhibit this property since it employs nonrobust sample estimates for computing higher order moments. In order to overcome this drawback, we propose to use robust alternatives to higher order moments, whic...

Introduction Most fMRI data display significant higher-order statistics representing tendencies to grouping along various shapes, even if such feature is commonly hidden by the overall distribution. Since the spatiotemporal characteristics of brain activity are frequently unknown and variable, their evaluation using hypothesis-driven methods only is rather difficult. Analysis of an fMRI block-b...

This paper reports the first results on extracting a meaningful representation for words from multilingual parallel corpora. Independent component analysis is used to extract a number of components from statistics calculated for words in contexts. Individual components are meaningful and multilingual and words are represented as a bag of concepts model. The component space created by the extrac...

Andrea Mazzolini, Marco Gherardi, 3 Michele Caselle, Marco Cosentino Lagomarsino, 3, 4 and Matteo Osella∗1 Physics Department and INFN, University of Turin, via P. Giuria 1, 10125 Turin, Italy Sorbonne Universités, UPMC Univ Paris 06, UMR 7238, Computational and Quantitative Biology, 15 rue de l’École de Médecine Paris, France CNRS, UMR 7238, Paris, France FIRC Institute of Molecular Oncology (...