PENGUKURAN PERFORMA SUPPORT VECTOR MACHINE DAN NEURAL NETWOK DALAM MERAMALKAN TINGKAT CURAH HUJAN

Abstract

Prediction models using Neural Network or Support Vector Machine have been developed in many areas of rainfall. In this studi we have compared the performance of NN and SVM models, in predition of monthly rainfall at R-17 station Kecepit Pemalang. The analyzed using NN and SVM in which testing with Root Mean Square Error (RMSE) for get the performance is done. The data obtained for 2009 to 2018 monthly rainfall were used as modelling and forecasting sample. The results showed that NN obtained smallest error rate compared to SVM. the recognized value of RMSE for SVM is 176,374, Neural Network is 22,289. In RMSE the smallest error rate showed the best performance of algorithm.