ABSTRACT
The ship classification project involved recognition of six ship classes through their radar range profiles. The data were gathered by an airplane flying in large circles around ships of six distinct structural types. The data consisted of a number of profiles for each of the different ship classes taken at angles about 20° apart around the compass. The goal was to construct a classifier which could take as input a profile at an unknown angle from one of the six classes and produce reliable predictions of class membership. Tree structured classifiers, or, more correctly, binary tree structured classifiers, are constructed by repeated splits of subsets of X into two descendant subsets, beginning with X itself. The first problem in tree construction is how to use L to determine the binary splits of X into smaller and smaller pieces.
