ABSTRACT
In may applications, data evolve with time. This is especially typical for relational data. Evolutionary clustering on relational data is extremely challenging, since it involves a complicated system of multiple types of data objects and relations changing over time. In this book, we start with general evolutionary clustering as an initial efforts for evolutionary clustering on relational data. In evolutionary clustering, we need to address two issues at each time step: the current clustering pattern should depend mainly on the current data features; on the other hand, the current clustering pattern should not deviate dramatically from the most recent history, i.e., we expect a certain level of temporal smoothness between clusters in successive time steps. In this chapter, we propose three evolutionary models based on the recent literature on Hierarchical Dirichlet Process (HDP) and Hidden Markov Model (HMM). Those models substantially advance the literature on evolutionary clustering in the sense that not only they both perform better than the existing literature, but more importantly they are capable of automatically learning the cluster numbers and structures during the evolution (for more detailed description of the models refer to [131,132]).
