"Data Mining and Knowledge Discovery in Databases Using Decision Trees"
Umair Aziz 2000
Abstract
Recently, there has been a great deal of interest in Data Mining and Knowledge Discovery in Databases (KDD). Decision trees, as a method for classification and regression, are of particular importance as they utilize symbolic and interpretable representations. This paper describes the use of decision trees in data mining applications and analyzes the KDD process by discussing some popular decision tree methods such as, ID3, CART, OC1, and CA. We find that OC1 is an efficient system that provides good oblique splits at each node of a decision tree. We present an empirical study using real data on these algorithms and integrate the results as a part of the overall KDD process.