Prediction performance of improved decision tree-based algorithms: a review

Applications of machine learning can be found in retail, banking, education, health sectors etc. To process the large data emanating from the various sectors, researchers are developing different algorithms using expertise from several fields and knowledge of existing algorithms. Machine learning decision tree algorithms which includes ID3, C4.5, C5.0, and CART (Classification and Regression Trees) are quite powerful. ID3 and C4.5 are mostly used in classification problems, and they are the focus of this research. C4.5 is an improved version of ID3 developed by Ross Quinlan. The prediction performance of these algorithms is very important. In this paper, the prediction performance of decision tree algorithms will be studied, an in-depth review will be conducted on relevant researches that attempted to improve the performance of the algorithms and the various methods used. Comparison will also be done between the various tree based algorithms. The major contribution of this review is to provide researchers with the progress made so far, as there is no available literature that has put together relevant improvements of decision tree based algorithms, and lastly lay the foundation for future research and improvements.