Abstract
The use of cluster analysis requires the active participation of the analyst to interpret the results and judge their significance. This stage of the process is subjective, intuitive, and heuristic. When entities bearing a previously unsuspected relationship are placed side by side as a result of clustering, their juxtaposition might be sufficient to spark the recognition or insight, which leads to discovery; clustering can relocate an entity from its customary context. This chapter describes some techniques that might aid the analyst and enhance his powers of discovery. These techniques are simple. The mechanical results of a hierarchical clustering algorithm can be described by reporting for each of the n–1 stages the identity of the clusters merged at that stage and the value of the criterion characterizing this merge.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
