Spatial temporal analysis of apartment hotspot detection

A Research in Constraint-Based Clustering

1. What are the differences between supervised clustering and supervised classification?

2. We proposed an efficient method that clusters hotspot houses by taking obstacles as constraints, such as gardens and forest. In many cases, one may like to cluster hotspot houses by considering high way-distance and the (specified) speed on the highway road. For example, two nearby houses without close-by highway access should not be considered close, and hotspot houses linked by 20 minutes local, slow highway should be considered even "farther" than those linked by 10-minutes highway roads. Can you work out an efficient clustering algorithm under such constraints?

3. Why does the researcher guidance often lead to meaningful spatial clustering? Suppose you want to build a user-guided clustering mechanism for On-line shopping.