Optimasi Parameter Holt–Winters Exponential Smoothing Menggunakan Multivariabel Golden Section Untuk Prediksi Penjualan Mobil Indonesia

Mamluatul Hani'ah (Politeknik Negeri Malang)
Yogi Kurniawan (Universitas Brawijaya Malang)

G-Tech : Jurnal Teknologi Terapan
Vol 7 No 2 (2023): G-Tech, Vol. 7 No. 2 April 2023

Automotive companies play a crucial role in absorbing labor in Indonesia, making their sustainability essential for the country. To ensure their sustainability, sales forecasting is vital as it can estimate production volume. Although Holt-Winters exponential smoothing is a reliable method for sales forecasting, it heavily depends on the accurate determination of α, β, and γ parameters. Inaccurate parameter determination can lead to inaccurate predictions, but this issue can be addressed by optimizing parameters. To this end, this study proposes integrating Holt-Winters exponential smoothing with multivariable golden section to generate optimal parameters for forecasting car sales in Indonesia. The proposed method yields an average MAPE of 9%, which is highly accurate in the forecasting scale. Furthermore, the proposed method outperforms previous studies in the same domain.