Demand forecasting models used in airline revenue management are primarily based on airline’s own sales data. These models have limited visibility into overall market conditions and competitive landscape. Market factors significantly influence customer behavior and hence should be considered for determining optimal control policy. We discuss data sources available to airlines that provide visibility into the future competitive schedule, market size forecast, and market share estimation. We also describe methodologies based on Machine Learning algorithms that can use to forecast these quantities and explain how they can be leveraged to improve pricing and revenue management practices.

航空公司收入管理中使用的需求预测模型主要基于航空公司自己的销售数据。这些模型对整体市场状况和竞争格局的可见性有限。市场因素显着影响客户的行为，因此应考虑确定最佳控制策略。我们讨论了可供航空公司使用的数据源，这些数据源提供了有关未来竞争时间表，市场规模预测和市场份额估计的可见性。我们还基于机器学习算法描述了可用于预测这些数量的方法，并解释了如何利用它们来改进定价和收益管理实践。