Font Size:
PERAMALAN JUMLAH PRODUKSI PANEN KELAPA SAWIT DENGAN MENGGUNAKAN METODE EXPONENTIAL SMOOTHING HOLT-WINTERS
Last modified: 2019-10-25
Abstract
The number of palm oil being harvested is the most important factor in order to determine specific policies in PT. Subur Agro Makmur. To overcome the existing problems, PT. Subur Agro Makmur requires forecasting for future production based on harvest experience in previous years. In its implementation, this study uses monthly harvest data which are grouped on an annual basis with continued group division based on land blocks. Each block will be calculated separately so that the calculation on one block will not affect the calculation on the other block. The steps taken consisted of identifying problems, collecting data, analyzing data, and testing the quality of forecasting results based on errors that occurred in the forecasting results using the MAD, MSE, and MAPE error analysis models. In the process of data analysis, the method used is exponential smoothing holt-winters with Additive models. The results showed that the exponential smoothing holt-winters method can be used to predict yields with MAD-based error rates reaching 33.56, and the error rate based on MSE calculations reached 2110.68, and resulted in an error rate of 3.96% on a calculation basis MAPE with a weighting coefficient for α = 0.9, β = 0.2 and γ = 0.9.