A1032
Title: Weighting games in clustering
Authors: Marianthi Markatou - University at Buffalo (United States) [presenting]
Abstract: Conceptual and technical challenges are investigated in variable weighting and selection in cluster analysis of mixed-type data. The specific objectives of existing weighting procedures are quite different and are often selected without a discussion of competing objectives and the implications of these selection choices. The different objectives of existing weighting and variable selection techniques are described and investigated, and then the variable weighting technique MEDEA (Multivariate Eigenvalue Decomposition Error Analysis) is developed. It is shown that MEDEA weighting fills an important and previously unfilled niche in cluster analysis of mixed-type data. Its performance in numerous experiments with both synthetic and actual datasets is demonstrated.
