Prediction of East Java province monthly sugar prices with a nonparametric regression and ARIMA

The objective of this study was to predict monthly sugar prices in East Java province. Increasing population from Indonesia, reaching 275 million people by 2024, heightens the urgency to address the growing demand and need for staple goods, including sugar. To achieve this, a nonparametric regression analysis method, namely the local polynomial estimator, was employed. In addition to this method, A technique for time series analysis using Autoregressive Integrated Moving Average was also used to determine the most appropriate method. The aim of this research was to determine the optimal monthly sugar price model and forecasting methodology for East Java province from May 2019 to April 2024, using a total of 60 data points. The data exhibited a fluctuating value, with the smallest value of 11,294 and the largest value of 17,911. The results indicate that the in-sample, out-sample, and all-sample estimations, conducted using a second-order local polynomial estimator, produced MSE values of 36,635.61, 2,145.734, and 31,053.42, R-square values of 97.65%, 99.96%, and 98.71%, as well as MAPE values of 0.66%, 0.34%, along with 0.61%, correspondingly. These results emphasize the great precision and dependability of the second-order local polynomial estimator, especially when compared to a time series analysis method, which resulted in a significantly higher MAPE value of 11.54%.