Interpretable Airfare Prediction Using Random Forests: Seasonal Variation, Statistical Testing, and a Structured Prediction Workflow

Airfare prediction remains a significant challenge in transportation economics, as fares are influenced by a combination of temporal, operational, and market-related factors, resulting in complex pricing dynamics. This study proposes a more interpretable datadriven framework that ties together statistical checking with ensemble machine learning to investigate fare variations. This study uses the US Airline Flight Routes and Fares (1993–2024) dataset from the Bureau of Transportation Statistics, followed by preprocessing, feature engineering, and inferential testing to examine the most influential factors for airfare dynamics. First, seasonal variation was evaluated using formal statistical tests to determine whether the quarterly differences in average airfare were statistically significant, thereby justifying the inclusion in the predictive model. The outcomes from this study show noticeable seasonal effects, with a one-way analysis of variance finding statistically significant variation across quarters (F = 46.19, p < 0.001). In addition, a structured prediction workflow has been built, so the system can accept user-specified inputs and return scenario-based airfare forecasts using the trained model. Overall, the framework offers a clear, usable approach for understanding seasonal shifts and operational patterns in airfare pricing. Received: 28 November 2025 | Revised: 6 May 2026 | Accepted: 29 September 2026 Conflicts of Interest The authors declare that they have no conflicts of interest to this work. Data Availability Statement Data sharing is not applicable to this article as no new datasets were generated for the study. The analysis was conducted using publicly available data from the open-source repository at US Airline Flight Routes and Fares 1993-2024. The source code supporting the findings of this study is available in RaYDaR at https://doi.org/10.25421/yorksj.32509494. Author Contribution Statement Swathi Ganesan: Conceptualization, Methodology, Software, Validation, Data curation, Writing – review & editing, Supervision, Project administration, Funding acquisition. Amjad Alam: Software, Validation, Funding acquisition. Sangita Pokhrel: Validation, Data curation. Ganapathy Kumar: Data curation, Supervision, Project administration. Shiv Kunwar: Software, Resources, Writing – original draft. Minel Somasiri: Formal analysis, Resources, Writing – original draft. Nalinda Somasiri: Formal analysis, Supervision, Project administration.