Date of Publication :31st December 2017

Abstract: A numerous amount of data is generated in the healthcare sector and mining of those data yields good results. It will also be helpful in diagnosing a particular disorder or a disease and also helps in predicting the future of the disease occurrence. Various classification algorithms in data mining have been used in research to predict a particular disease. Classification is used to find out in which group each data instance is related within a given datas et. It classifies the data into different classes based on some conditions There are many classification algorithms which includes C4.5, ID3, k-nearest neighbor, Naive Bayes, SVM, and ANN etc., which are used for classification. There are three forms of classification approaches, namely Statistics, Machine Learning and Neural Network for classification. The main objective of this study is to provide a compact source of reference for the researchers who want to use decision tree which is an important tool of data mining technology in their area of work. With this aim in mind, we compared widely used decision tree algorithms to classify types of disease and compared their performances according to six performance metrics (ACC(%), MAE, PRE, REC, FME, and Kappa Statistic). We hope that this study can provide a useful overview of the current work in this field and highlight how to apply decision tree algorithms as a tool of data mining technology. While considering these approaches this paper provides an inclusive survey of different classification algorithms of decision trees and their features and limitations.

Reference :