MULTIVARIATE SINGULAR SPECTRUM ANALYSIS MODEL IN FORECASTING RED CHILI AND CAYENNE PEPPER PRICES

Radita Rahma (Universitas Islam Negeri Sumatera Utara)
Rina Filia Sari (Universitas Islam Negeri Sumatera Utara)
Sajaratud Dur (Universitas Islam Negeri Sumatera Utara)

AL-ULUM: JURNAL SAINS DAN TEKNOLOGI
Vol 10, No 1 (2024)

North Sumatra is one of the provinces that contributes the most prominent agricultural commodities of red chilies and cayenne peppers in Indonesia. This study aims to determine the outcomes of price forecasts for cayenne and red chilies in the province of North Sumatra. The method used is multivariate singular spectrum analysis. Results were grouped into six groups based on 12 eigenvectors with a forecast length of 9 monthly periods. Further, the level of accuracy was obtained from MAPE for each variable, with the highest MAPE being 30% for the curly red chili, 27.55% for the big red chili, and the lowest at 23.44% for mixed cayenne pepper. So, the price forecast for red chilies and cayenne peppers in North Sumatra Province for October 2023 to June 2024 using the Multivariate Singular Spectrum Analysis model is included in the forecast category with reasonable capabilities.