Passenger purchase patience assessment and willingness to pay prediction based on online advanced purchase behaviors data

This article develops a predictive framework using online advance purchase behavior data to estimate civil aviation passengers’ willingness to pay (WTP). The framework introduces a passenger purchase patience (PPP) model to quantify passengers’ tolerance of time cost and decision-making uncertainty during ticket acquisition. Based on PPP indicators, K-means clustering is used to classify passengers into low-, medium-, and high-patience groups. To predict purchase probabilities across these groups, we propose a decaying random survival forest (DRSF) algorithm that dynamically adjusts sample weights over the sales period. Passenger WTP is then estimated through a back-calculation formula based on predicted purchase probability. Experimental results show that purchase probabilities vary systematically by patience level as departure approaches. Low-patience passengers are less price-sensitive, medium-patience passengers respond strongly to prices below WTP, and high-patience passengers show stronger planning commitment and price resistance.