Research on Consumer Hotel Booking Channel Selection Based on Machine Learning

• 1、School of Economics and Management, Beijing University of Posts and Telecommunications，Beijing 100876

Abstract：This paper uses a 5-fold cross-validation recursive feature elimination algorithm based on the extreme gradient lifting method to mine a resort hotel data set. It is found that three factors which have a major impact on consumers' choice of hotel reservation channels are mainly whether to find agents, average daily rate, and consumer type. It has been proved from the data level that consumer characteristics and consumption patterns have the most important effect on the choice of consumer hotel reservation channels. Consumers who usually have agents participating in bookings usually choose third-party agent channels. And those who plan at a long time before his arriving, he would have more possibility to choose a third-party agency channel, while the more demanding consumers tend to choose a hotel's direct sales channel. The comparison of the performance of the four classification methods before and after verifies the effectiveness of feature filtering. The logistic regression model is used to measure the impact of features from an objective level. Conclusions are combined with theories such as service quality model, time construction theory, perceived price and so on. It explained that the hotel can conduct omni-channel construction by scientifically managing the relationship with the online travel agency to improve hotel revenue and efficiency.

Keywords： information management and information system hotel booking channel XGBoost online travel agency recursive feature elimination