Komparasi Support Vector Machine (Svm) Dan Autoregressive Integrated Moving Average (Arima) Pada Peramalan Hujan Di Daerah Albury, Australia

Abstract

Rain is one of the 4 weather conditions on earth. The rain itself is sometimes erratic. For some people, rain is sometimes seen as an obstacle in carrying out their daily activities. Therefore, predictions about rain are needed. Predictions about rain really help people in carrying out their daily activities. The research aims to compare the predictive performance of SVM and ARIMA in forecasting rain. The data used in this case study research is data regarding daily weather for 10 years in the Albury area, Australia from 2008-2017 with a lot of 3040 data. The results obtained from SVM for forecasting rain using daily weather data for 10 years in the Albury area, Australia with the best accuracy rate with the SVM model is 97.532% with an error rate of 2.468%, while in the ARIMA model, the results are MAE 0.181, RMSE 0.254 and MAPE 0.159. So it can be concluded that the ARIMA model has a better performance in predicting rain than the SVM method.

Penerapan Metode Support Vector Machine Dalam Memprediksi Prediksi Cuaca

“…Bahkan ada pihak yang menginginkan tersedianya prakiraan atau ramalan cuaca dalam rentang waktu yang cukup kecil misalnya waktu harian, jam, dan bahkan dalam waktu menit. Jadi dengan adanya kebutuhan seperti ini, mendorong peneliti untuk terus melakukan penelitian-penelitian terkait atmosfer cuaca dan juga mengembangkan metode untuk menentukan cuaca sehingga mampu mendapatkan hasil yang baik dan akurat [4], [5] [6]. Prakiraaan dapat dilakukan dengan berbagai macam cara, salah satunya dengan menggunakan metode data mining.…”