Rice Price Prediction in Province Nusa Tenggara Barat (NTB) Using Comparison Of Linear Regression And Random Forest Algorithms

Abstract

This research provides transparent prediction accuracy in rice industry management and can be used to more accurately forecast prices in the West Nusa Tenggara region. To determine whether the model provides better prediction accuracy, the researchers analyzed and forecasted rice prices using two machine learning algorithms: Linear Regression and Random Forest. Forecasting rice prices is difficult because the elements that support changes in rice prices, such as planted land, production levels, consumption levels, currency (rupiah) volatility, and the volume of rice imports into Indonesia, are interrelated. Based on the study, Random Forest outperformed Linear Regression, with an R² value of 0.710, indicating a better model fit. In addition, the Random Forest algorithm shows a lower error rate, which is reflected in the RMSE of 1038,394. The dataset used for this study covers the period 2006 to 2021 and is sourced from various official institutions, including the Central Bureau of Statistics and Bank Indonesia