Comparison Of Extreme Learning Machine And Holt Winter’s Exponential Smoothing Methods In Railway Passenger Forecasting

Abstrak

Trains are public transportation consisting of a locomotive pulled by other carriages as its main part. The carriages are very large so that they can transport passengers and goods on a large scale. Forecasting is a process to estimate how much future demand will be which includes the needs in terms of quantity, quality, time, and location needed to meet the demand for goods or services. The method used in forecasting this problem is Holt Winter Exponential Smoothing. The Holt Winter method is a moving average forecasting method that gives weight to past data exponentially so that current data has a greater weight in a moving average that uses 3 levels of smoothing. This method is used when the data shows trends and seasonal. The three smoothing parameters are α (for the process level), β (for the trend element), and γ (for the seasonality element) with values ​​between 0 and 1 for each parameter. Extreme Learning Machine (ELM) is a simple feedforward artificial neural network development method using one hidden layer or commonly known as a single hidden layer feedforward neural network. The data used is the number of train passengers from January 2021 to December 2023. The results of the study showed that the best model in modeling the number of train passengers was Holt Winters Exponential smoothing. The accuracy of the model is good with a MAPE value of 17.10% compared to ELM 20%.