A new approach for classification of patterns having categorical attributes

This paper proposes a novel approach for classification of patterns having categorical attributes using decision trees. A new split measure has been proposed for construction of decision trees. Main focus of the proposed split measure is to improve the classification accuracy. Performance of the proposed split measure has been compared with the well known split measure information gain used in ID3 algorithm. It has been shown that the proposed split measure outperforms information gain on benchmark datasets taken from UCI machine learning repository.