Penerapan Algoritma Support Vector Regression dalam Memprediksi Produksi dan Produktivitas Kelapa Sawit

Abstract

Palm oil is a plantation crop that provides the highest economic value in Indonesia. Riau is currently the highest palm oil producing province in Indonesia with a state-run palm oil company, PTPN V. However, palm oil production is not always stable every month, whichexperiences ups and downs in the amount of production and productivity due to several factors including irregular rainfall, climate, soil fertility and most importantly fruit bunches that are not ready to harvest. So the data mining processing process is carried out by predicting the amount of production and productivity of oil palm applying the Support Vector Regression (SVR) algorithm with three kernels such as the Linear kernel, RBF kernel and Polynomial kernel. Experimental results on palm oil production and productivity show that the best kernel is the RBF kernel because the prediction results are close to the actual value. The accurate rate on palm oil production is 75.4% and palm oil productivity produces an accuracy value of 71%. It also produces an error value on palm oil production of 1.8%, for productivity of 2.1%. The results of the study can be used as an estimated picture in the company's future decision making.