El Centro de Investigación en Matemáticas, A.C. (Unidad Monterrey), se complace en invitarlos a la conferencia Clustering Binary Data by Application of Combinatorial Optimization Heuristics, del Dr. Javier Trejos-Zelaya* (Universidad de Costa Rica), el jueves 24 de septiembre a las 12:00 horas, a realizarse de manera virtual a través del siguiente enlace: https://bluejeans.com/469167172
Resumen:
We study clustering methods for binary, or 0/1, data. This kind of data require the definition of aggregation criteria that measure the compactness of each cluster; we study some of these criteria and deduce some theoretical properties. Five new and original methods for clustering binary data are introduced, using neighborhoods and population behavior combinatorial optimization metaheuristics: first ones are simulated annealing, threshold accepting and tabu search, and the others are a genetic algorithm and ant colony optimization. The methods are compared to classical ones, such as hierarchical clustering, and two versions of k-means: dynamical clusters and partitioning around medoids or PAM. The methods are implemented, performing the proper calibration parameters in the case of heuristics, to ensure good results.
From a set of 16 data tables generated by a quasi-Monte Carlo experiment, a comparison of the results obtained by classifying the objects in each data tab
* El Dr. Javier Trejos Zelaya es Licenciado en Matemática por la Universidad de Costa Rica
(1988) y Doctor en Matemática Aplicada por la Universidad Paul Sabatier de Toulouse, Francia
(1994). Catedrático de la Escuela de Matemática, fundador del Centro de Investigación en
Matemática Pura y Aplicada (CIMPA) en 1997.