Comparative analysis of time series forecasting for Nepal Airlines passenger data: ARIMA vs. LSTM model

Abstract

This paper presents a comparative study of time series forecasting methods applied to Nepal Airlines passenger data, focusing on the ARIMA and LSTM models. The study aims to analyze the forecasting performance of these models and identify the most accurate approach for predicting future airline passenger numbers. The ARIMA model captures linear trends and seasonality, while the LSTM neural network is employed for its ability to model complex patterns and non-linear relationships within the data. Both models are evaluated using standard performance metrics, and the results provide insights into the strengths and weaknesses of each forecasting technique. The results indicate that ARIMA provided more accurate forecasts with MAE: 0.74 and RMSE: 1.78, compared to LSTM having MAE: 0.87 and RMSE: 2.02, underscoring its suitability for datasets with linear trends and seasonality.