OTPS: A decision support service for optimal airfare Ticket Purchase

Due to the fast development and wide application of Web technology, online purchases have become the main method to buy airfare tickets. Because information about pricing factors is unavailable, it is difficult for customers to buy airfare tickets at the lowest cost or even at a relatively lower cost. Although several approaches have been proposed to suggest the optimal timing for purchasing airfare tickets, these methods are based on predictions of tomorrow's price information. Clearly, local optimal decisions which rely solely on tomorrow's price prediction can be misleading or incorrect, resulting in missed opportunities to buy tickets at a lower cost in the future. In this paper, a novel Optimal airfare Ticket Purchase decision-support Service (OTPS) is proposed, which makes continuous recommendations on the best purchase time before the departure date. OTPS is based on the strategy of the Dynamic ratio of numbers of Potential days with Lower Price (DPLP), which takes the fluctuation tendency of airfare prices for a period into consideration. Specifically, parameter values of the model are set in a route-specific manner and updated periodically to enhance the generalization ability and reliability of OTPS. Extensive experiments are conducted on the real-world ticket price dataset of multiple routes. The experimental results prove that OTPS shows superiority over other state-of-the-art solutions.