Building a Classifier Employing Prism Algorithm with Fuzzy Logic

Abstract

Classification in data mining is receiving immense interest in recent times. As the knowledge is based on historical data, classifications of data are essential for discovering the knowledge. To decrease the classification complexity, the quantitative attributes of data need splitting. But the splitting using the classical logic is less accurate. This can be overcome by the use of fuzzy logic. This paper illustrates how to build up the classification rules using the fuzzy logic. The fuzzy classifier is built on by using the prism decision tree algorithm. This classifier produces more realistic results than the classical one. The effectiveness of this method is justified over a sample dataset.