Komparasi Pergerakan Harga Nikel dengan Metode Double Exsponensial Smoothing, Trend Kuadratik, dan Trend Eksponensial

Abstract

Nickel is a crucial mineral commodity in the manufacturing industry and the global energy transition, particularly as a raw material for electric vehicles. The volatility of nickel prices creates a need for forecasting methods capable of generating accurate price estimates. This study aims to analyze and compare the accuracy of three time-series forecasting methods—Double Exponential Smoothing (DES), Quadratic Trend, and Exponential Trend—in projecting nickel prices. The research uses a quantitative approach with 30 observations of monthly nickel price data for the period January 2022 to June 2024 obtained from the public, namely Investing.com and related publications. The accuracy of each method was evaluated using Mean Absolute Deviation (MAD), Mean Squared Error (MSE), and Mean Absolute Percentage Error (MAPE). The results indicate that the Double Exponential Smoothing method achieved the highest accuracy, with a MAPE of 8.45%, whereas the Quadratic Trend and Exponential Trend methods yielded MAPE values ​​of 11.47% and 11.89%, respectively. Based on the best-performing method, the projected nickel price for July 2024 is 20,423.40 USD/dmt. This study contributes an empirical comparison of the effectiveness of these three forecasting methods regarding fluctuating monthly nickel price data and provides a basis for selecting more accurate forecasting methods to support planning and decision-making related to nickel prices.