FORECASTING CATEGORY IN PUBLIC SCHOOL STAFF STUDENTS AND SPENDING IN 2018 WITH COMPARISON METHOD

Education is one of the efforts to develop capabilities and shape the character and civilization of a dignified nation. Increasingly expensive education costs require parents to prepare more maturely. Therefore, the purpose of this study is to show forecasting of rill costs per student in 2018 using the Timeseries Public School Staff, Students, and Spending 1970-2017 data. In this study, the ARIMA and the SVM model are used to forecast rill costs per student in 2018. ARIMA is one of the simplest models and the most widely used method. This model will be compared using machine learning, namely SVM. From the experimental results that compare the performance of SVM and ARIMA. In the calculation above, the MAPE results from the Arima and SVM models are 0.142 and 0.0029. Whereas for RMSE the results of the ARIMA calculations obtained results of 36.625, while for SVM the RMSE results obtained were 7.470. The results of the study concluded that the SVM model has a better performance in predicting crude oil prices compared to the ARIMA method.