PENERAPAN METODE SINGLE EXPONENTIAL SMOOTHING UNTUK MEMPREDIKSI JUMLAH PENUMPANG KERETA API

Abstract

Abstrak

This study aims to apply the Single Exponential Smoothing method to predict the number of train passengers in

Indonesia to support the operational planning of PT Kereta Api Indonesia (KAI). The modernization of

infrastructure and improvement of train services have been a top priority for the government to enhance public

mobility and economic growth. However, fluctuations in the number of passengers affected by seasons, holidays,

and special events pose a challenge in managing railway transportation. The Single Exponential Smoothing

method was chosen due to its ability to assign greater weight to more recent data, allowing for quick capture of

trend changes and providing more accurate predictions. This study uses historical passenger data to analyze and

model forecasting using various alpha values to determine the best level of accuracy. The results show that using

alpha values of 0.1, 0.3, 0.5, and 0.8, the best MAD measurements are achieved with Jabodetabek at alpha=0.3

(1052), Non-Jabodetabek at alpha=0.5 (586), Java at alpha=0.8 (1235), Sumatera at alpha=0.1 (29), and Airport at

alpha=0.5 (38). The best MSE measurements are Jabodetabek at alpha=0.1 (1,953,420), Non-Jabodetabek at

alpha=0.3 (651,336), Java at alpha=0.8 (2,521,278), Sumatera at alpha=0.1 (1487), and Airport at alpha=0.1

(2306). Meanwhile, the best MAPE measurements are Jabodetabek at alpha=0.3 (3.90%), Non-Jabodetabek at

alpha=0.5 (8.34%), Java at alpha=0.8 (3.68%), Sumatera at alpha=0.1 (6.44%), and Airport at alpha=0.3 (5.74%).