Authors: Mohammad Reza Afrash; Mehdi Khalili; Maral Sedigh Salekde
Addresses: Department of Computer and Informatics Engineering, Payame Noor University, Tehran, Iran ' Department of Computer and Informatics Engineering, Payame Noor University, Tehran, Iran ' Department of Computer and Informatics Engineering, Payame Noor University, Tehran, Iran
Abstract: Heart disease is a term that covers a range of disorders that affect heart. Medical diagnosis is a very important and it is a complicated task that should be performed properly and efficiently. The needs for new tools able to help doctors in predicting and diagnosis heart disease which is highly recognised. In the present study first, the dataset containing total instances of 439 with 14 attributes were obtained from several medical centres in Tehran, Iran. Secondly, five data mining techniques were used: the naïve Bayes, ANN, random forest and C4.5 decision trees and random tree algorithms. From the experimental result it is observed that the Random forest algorithm with (94.53, 0.945, 0.969, 0.945, 0.916) for accuracy, F-measure, specificity, sensitivity and kappa rate produce a higher performance for our classification model when compared with other algorithms.
Keywords: data mining techniques; heart disease; classification; weak.
International Journal of Advanced Intelligence Paradigms, 2020 Vol.16 No.1, pp.88 - 97
Received: 21 Mar 2017
Accepted: 02 Jun 2017
Published online: 20 Apr 2020 *