PERAMALAN PRODUKSI KOPI DENGAN MENGGUNAKAN METODE TIME SERIES DI KABUPATEN BATANG, JAWA TENGAH
Kata Kunci:
Coffee, Forecasting, Time Series, Moving Average, Exponential Smoothing, Linear Regression, kopi, Peramalan, Regresi LinearAbstrak
Coffee is one of the important plantation commodities in Batang Regency, making a significant contribution to its economic development, facing fluctuating production patterns that are influenced by the dynamics of industrial policies and strategies that are implemented. To overcome this challenge, the use of time series forecasting in production forecasting is imperative. A comprehensive study was conducted to project 2023 coffee production using three different methods. In particular, the moving average (MA) method produces a projection of 756,891 kg, accompanied by metric errors (MAD = 56,348.88, MSE = 5,031,788,555.3, MAPE = 7.93%). Likewise, the linear regression (LR) method anticipates 883,743.73 kg, along with the associated error (MAD = 91,495.81, MSE = 16,004,246,807.16, MAPE = 20.24%), while the exponential smoothing (ES) method predicts 725,499 .74 kg, with a value-based error (MAD = 91,904.05, MSE = 30,840,632,820.92, MAPE = 12.26%). In comparison, the moving average (MA) method exhibits exceptional accuracy, resulting in minimal error (MAD = 56,348.88, MSE = 5,031,788,555.3, MAPE = 7.93%), making it a superior forecasting technique. and most precise. Keywords: Coffee, Forecasting, Time Series, Moving Average, Exponential Smoothing, Linear Regression.
Kopi merupakan salah satu hasil perkebunan di Kabupaten Batang yang berperan penting dalam menunjang pembangunan ekonomi Kabupaten Batang. Namun produksi kopi di wilayah Kabupaten Batang menunjukkan kecenderungan yang fluktuatif yang disebabkan oleh faktor kebijakan dan strategi produksi yang diterapkan di industri kopi. Peramalan produksi dengan metode time series digunakan untuk meramalkan kebutuhan produksi. Hasil dugaan produksi kopi tahun 2023 dengan menggunakan tiga metode adalah sebagai berikut: Metode moving average (MA) menghasilkan 756.891 kg dengan nilai error MAD = 56.348,88, UMK = 5.031.788.555,3 dan MAPE = 7,93%. Sedangkan metode regresi linier (LR) memprediksi 883.743,73 kg dengan nilai error MAD = 91.495,81, MSE = 16.004.246.807,16 dan MAPE = 20,24%, serta metode exponential smoothing (ES) memprediksi 725.499,74 kg dengan nilai error MAD = 91.904,05, MSE = 30.840.632.820,92 dan MAPE = 12,26%.Metode moving average (MA) terbukti sebagai metode peramalan yang paling baik dan akurat, menghasilkan nilai error yang minimal dengan MAD = 56.348,88, MSE = 5.031.788.555,3 dan MAPE = 7,93%. Kata kunci: kopi, Peramalan, Time Series, Moving Average, Regresi Linear, Exponential Smoothing.
