Online Tourism Purchase Analysis and Prediction

With the rapid development of tourism e-commerce, a huge amount of online tourists behavioral data is enlarged at an explosive speed. Online purchase analysis by making full use of the behavioral data undoubtedly is crucial to achieve precision marketing. Along this line, this paper offers an empirical analysis on online purchase of tourism products, and thus attempts to construct a suite of effective features for the task of online purchase prediction. In particular, our analysis indicates several interesting characteristics of e-travel data that are clearly different with traditional online purchasing data, and also points out some key factors affecting the purchasing decisions of tourism products. Based on this analysis, we present twelve effective features for the subsequent purchase prediction task. Experimental results demonstrate that the proposed features can substantially improve the performance of the ensemble classifier for purchase prediction.