Tracing cluster transitions for different cluster types
Irene Ntoutsi, Myra Spiliopoulou, Yannis Theodoridis · 2009
Abstract: Clustering algorithms detect groups of similar population members, like customers, news or genes. In many clustering applications the observed population evolves and changes over time, subject to internal and external factors. Detecting and understanding changes is important for decision support. In this work, we present the MONIC + framework for cluster-type-specific transition modeling and detection. MONIC + encompasses a typification of clusters and cluster-type-specific transition indicators, by exploiting cluster topology and cluster statistics for the transition detection process. Our experiments on both synthetic and real datasets demonstrate the usefulness and applicability of our framework.