Webinar video https://www.youtube.com/watch?v=8ws_v3HRSD0
Presentation slides
Abstract
We propose sparse variants of correspondence analysis (CA) for large contingency tables like documents-terms matrices used in text mining. By seeking to obtain many zero coefficients, sparse CA remedies to the difficulty of interpreting CA results when the size of the table is large. Since CA is a double weighted PCA (for rows and columns) or a weighted generalized SVD, we adapt known sparse versions of these methods with specific developments to obtain orthogonal solutions and to tune the sparseness parameters. We distinguish two cases depending on whether sparseness is asked for both rows and columns, or only for one set.
Short Bio
Gilbert Saporta is since 2014 Emeritus Professor of Applied Statistics at Conservatoire National des Arts et Métiers, Paris, that he joined in 1984. He has been Guest Professor (2017-2022) at Beihang University, Beijing, China and taught in various universities.
After an engineering degree and a Master in Mathematical Statistics, he obtained a Ph.D. in 1975 and later a “Doctorat d’État” in 1981 both at Université Pierre et Marie Curie, (now Sorbonne Université), Paris. His field of research is multivariate data analysis, including categorical and mixed data analysis by means of optimal scaling techniques, correspondence analysis, Functional Data Analysis, supervised classification with applications in risk analysis. He has supervised 29 Ph.D thesis.
Gilbert has been a consulting statistician for various companies, e.g. L’Oréal and IPSOS.
He is honorary President of the French Statistical Society and has been President of IASC (International Association for Statistical Computing) and Vice-President of ISI (International Statistical Institute) 2005-2007.
Information at FEP News
More information at News U. Porto