ABSTRACT
This chapter is about the three important components of CPOV cluster analysis in static unlabeled data. The mismatch between AM and CM leaves a lot of room for misinterpretation of any algorithmic CPOV clusters discovered in the data during the partitioning step. The correct way to refer to CPOV clusters in data is to always identify them with reference to a particular model and algorithm. A source of confusion about cluster validity is due to the very fine line that may exist between the three problems of assessment, clustering, and validation. The canonical problems can be quite distinct for some approaches to cluster analysis but practically impossible to separate for others. Tryon paved the way for a different approach to visual clustering using visual assessment and aggregation of hand-rendered profile graphs for all canonical problems discussed in this chapter.
