Time-Based Clustering and Its Relationship with Mutual Information Theory

University of Utah

This paper re-introduces the concepts of TimeBased Clustering (TBC). Also, the ideas of oversampling and embedding time are introduced in connection with mutual information theory. These concepts are then extended through the use of the Time-Based Clustering (TBC) problem. Mutual information curves for the Rössler system are shown to match a slice through the rich cost function space spanned by the Time-Based Clustering (TBC) solution. In closing, some possible repercussions of this find are discussed.