Forecasting Airlines Customer Demand: A Comparative Analysis of Time Series Models

The airline industry is renowned for employing sophisticated strategies, making accurate customer demand fore-casting essential for operational planning and resource allocation. Revenue management systems heavily rely on precise passenger demand forecasts. This research paper consists of two main parts. Firstly, it compares and evaluates the performance of different time series models, namely SARIMA, ARIMA, and auto-regression, for forecasting airline customer demand. These forecasting methods are applied using features such as month, number of passengers, international flights, and total number of flights. Secondly, the models are tested on pattern analysis, including level, seasonality, and trend, and the evaluation metrics are assessed. The outcomes of this study contribute to the existing knowledge on forecasting airline customer demand, offering valuable insights for optimizing operational planning and resource allocation in the airline industry. The comparative analysis of forecasting models provides practical guidance for airlines to select the most suitable model for their specific requirements. The evaluation metrics, including Mean Absolute Error (MAE), Mean Squared Error (MSE), R-squared $\left(\mathbf{R}^2\right)$, and Mean Absolute Percentage Error (MAPE), clearly indicate the superior performance of the SARIMA methodology in terms of forecast bias. The experimental results demonstrate that SARIMA models incorporating seasonality patterns achieve the best forecasting outcomes. Keywords: Revenue management, accurate forecasts, evaluation matrix, time series algorithms