An Interval Classifier for Database Mining Applications

Abstract

We are given a large population database that contains information about population instances. The population is known to comprise of m groups, but the population instances are not labeled with the group identification. Also given is a population sample (much smaller than the population but representative ofit) in which the group labels of the instances are known. We present aninterval classifier (IC) which generates a classification function for each group that can be used to efficiently retrieve all instances of the specied group from the population database. To allow IC to be embedded in interactive loops to answer adhoc queries about attributes with missing values, IC has been designed to be efficient in the generation of classification functions. Preliminary experimental results indicate that IC not only has retrieval and classifier generation efficiency advantages, but also compares favorably in the classification accuracy with current tree classifirs, such as ID3, which were primarily designed for minimizing classification errors. We also describe some new applications that arise from encapsulating the classification capability in database systems and discuss extensions to IC for it to be used in these new application domains.