KOMPARASI ALGORITMA MULTI LAYER PERCEPTRON DAN RADIAL BASIS FUNCTION UNTUK DIAGNOSA PENYAKIT JANTUNG

Abstract

A neural network as a data mining model has many algorithms with different accuracy level. This research uses the UCI machine learning repository’s data to compare the accuracy level of Multilayer Perceptron (MLP) and Radial Basis Function (RBF) algorithm in predicting the heart disease. Confusion matrix and the ROC (Receiver Operating Characteristic) curve method is used to measure the performance of both algorithms. Based on the test results of the implementation, proved that the MLP algorithm has a higher value of accuracy than the RBF algorithm. Using the Confusion Matrix, the MLP algorithm has higher value of accuracy with 87.3%  than the RBF algorithm with 81.1%. Using the ROC curve, the MLP algorithm also has higher AUC (Area Under the Curve) value with 0.949 than the RBF algorithm with 0.911. Using confusion matrix, the accuracy value of both algorithms are included as good classification, because the AUC value is in the range of 0.80 until 0.90. Using ROC Curve, the accuracy values are included as excellent classification, because the AUC value is in the range of 0.90 until 1.00.