Decision tree algorithm is the fundamental and popular Data classification Algorithm. Iterative Dichotomiser (ID3) decision tree algorithm has achieved good results in most of the classification problems. But ID3 algorithm had some disadvantages such as attribute biasing and it reduces the classification Accuracy. Shannon Entropy measure is used in ID3 algorithm for classification. The motivation of the research work is to improve the performance of the classification Accuracy in decision tree generation and hence a novel approach to modify the Entropy with the Non parametric plug in Entropy Estimator based kernel equations is attempted. The proposed method proved good classification outcome for distribution of dataset irrespective of the complexity of the data sets. The relative comparison of the results based on classification accuracy is examined with traditional ID3 and our proposed Non parametric plug in Entropy estimators. The results with the proposed work performed well in all the experiments conducted.
Volume 12 | Issue 6
Pages: 1945-1958
DOI: 10.5373/JARDCS/V12I2/S20201400