Statistics and Risk Management Seminar - 18/3/2015

Wednesday, 18 March 2015, 3:00 p.m.

Lecturer: Susana Nascimento, Departamento de Informática and NOVA Laboratory for Computer Science and Informatics (NOVA-LINCS), FCT-UNL. (Joint work with Boris Mirkin, Computer Science Dep., Birkbeck University of London, UK, Dep. of Data Analysis and Artificial Intelligence, National Research University, Higher School of Economics, RF).

Title: "Data Recovery Fuzzy Clustering: an Additive Spectral Method and Applications"

Local: Sala de Seminários, Edifício VII
Faculdade de Ciências e Tecnologia, Quinta da Torre, Caparica

Abstract: In this talk we present an additive fuzzy clustering method for similarity data as oriented towards representation of activities of research organizations in a hierarchical taxonomy of its knowledge domain. We developed a one-by-one cluster extracting strategy which leads to an algorithm of spectral clustering for similarity data. The method, FADDIS, has been experimentally verified on simulated and real-world data. Specifically, we developed two synthetic data generators simulating affinity data of Gaussian clusters and genuine additive similarity data, with a controlled level of noise, and we discuss FADDIS results on these data in comparison with two state-of-the art fuzzy clustering methods. Then, we extend our analysis to “difficult” data structures taken from literature. Finally, we present FADDIS results when applied to in-house data of similarity between research topics according to the work of a research centre.