Partitioning data sets into disjoint groups is a problem arising in many domains. The theory of cluster analysis aims to find groups that are both homogeneous (entities in the same group that are similar) and well separated (entities in different groups that are dissimilar). There has been rapid expansion in the axiomatic foundations and the computational complexity of such problems and in the design and analysis of exact or heuristic algorithms to solve them. Applications have burgeoned in psychology, computer vision, target tracking, and other areas. This book contains papers presented at the workshop Partioning Data Sets held at DIMACS in April 1993. Some of the papers cover the main paradigms of the field of cluster analysis methods and algorithms. Other topics include partitioning problems arising from multitarget tracking and surveillance and from computer and human vision. The multiplicity of approaches, methods, problems, and algorithms make for lively and informative reading.