Forecasting the demand for railway freight transportation in Kazakhstan: A case study

Abstract

Effective resource planning in railway freight transportation necessitates precise demand forecasting, shaped by a complex array of dynamic variables. In Kazakhstan, the National State-owned Railway Company (KTZ) underwent a transition from traditional expert-based forecasting methods to the application of mathematical models, notably the Autoregressive Integrated Moving Average (ARIMA) model. Employing historical data spanning from 2012 to 2016, this shift signifies a significant evolution in KTZ's approach to anticipating freight demand.The primary objective of this study is the empirical assessment of the efficacy of the ARIMA model in comparison to conventional qualitative forecasting techniques. Through the analysis of actual data from 2017, utilizing established metrics such as the Mean Absolute Error (MAE) and the Mean Absolute Percentage Error (MAPE), this research substantiates the utility of time series analysis through ARIMA. The findings not only confirm the model's effectiveness but also emphasize its superiority in refining the precision of railway freight demand forecasts, particularly within the unique context of Kazakhstan.Beyond the validation of methodologies, this research serves as a precursor to advanced forecasting practices, offering the potential to redefine resource planning in the railway industry. By extending the predictive horizon to 2024, the manuscript aligns with contemporary standards, providing nuanced insights for operational and developmental considerations in Kazakhstan's railway freight sector. This expansion positions the study within the evolving landscape of the industry, ensuring a comprehensive and forward-looking contribution to efficient resource allocation and modernized planning practices.