A Data Mining Approach to Building a Predictive Model of Low-Cost Carriers' Presence in the U.S. Domestic Routes
The purpose of the study was to build the predictive model of the presence of U.S. low-cost carriers (LCCs) in the domestic network structure. SEMMA (Sample, Explore, Modify, Model, and Assess) schematic in data mining was followed and employed as the primary methodological procedure. Data in the period of 1Q2016-1Q2018 were extracted from the Bureau of Transportation Statistics (DB1B database) and reconstructed to form predictors. Stepwise logistic regression showed a significant predictive performance compared to decision tree technique in terms of fitting measures, which was then used as the concluding model. Significant predictors included: (1) Market concentration positively related with the presence of LCCs, (2) nonstop route associated with the presence of LCCs, (3) market airfare factors negatively related with the presence of LCCs, and (4) origin and destination (O&D) airports being hubs, especially medium hubs, associated with the presence of LCCs. The findings may practically aid network planners in airlines and airports in decision making associated with the presence of LCCs, which ultimately leads to building their more robust and efficient route map.
Scholarly Commons Citation
Deaton, J. E.,
A Data Mining Approach to Building a Predictive Model of Low-Cost Carriers' Presence in the U.S. Domestic Routes.
International Journal of Aviation, Aeronautics, and Aerospace,
Business Administration, Management, and Operations Commons, Business Analytics Commons, Management Sciences and Quantitative Methods Commons, Operations and Supply Chain Management Commons