Examining the Forecasting Movement of Palm Oil Price Using RBFNN-2SATRA Metaheuristic Algorithms for Logic Mining

Abstract

RBFNN with different algorithms and the logic mining method for forecasting constitute the most significant tools and techniques, which are used to demonstrate the economic growth in the country. Upon using monthly data spanning from Jan 2016 to March 2020 for the manufacturing of palm oil, the results mainly revealed that RBFNN-2SATRAAIS is the most accurate and efficient model compared to RBFNN-2SATRAPSO and RBFNN-2SATRAGA in forecasting the price of palm oil. RBFNN-2SATRAAIS had the highest average overall accuracy (90.476190%), followed by RBFNN-2SATRAPSO (85.71%), and RBFNN-2SATRAGA (76.19%). The results also showed that the spot price of palm oil is highly influenced by the total exports and imports, as well as the production of palm oil, its end-stocks, and Malaysia's real, effective exchange rate. By using the data mining technique based on the energy minimization technique, the logical mining task was carried out. The empirical findings provided useful insights into decision-making and policy implementations, including the formulation of strategies to help the industry in dealing with constant price fluctuations and, thereby, enabling the Malaysian palm oil industry to continue its domination over the international market.