Research about User Purchasing Behavior Analysis Based on Graph Convolutional Network and Web Crawler

In view of the lack of functional software oriented to “low price” in the current market, this paper recommends air tickets, hotels and appropriate tourist cities, and uses crawlers to capture and collect data on Ctrip and Qunar's air tickets, hotels and scenic spots, and then extracts their key features through graph convolutional neural network. Through the analysis of the obtained data and information, to achieve the season special recommendation and popular attractions recommendation. It fills the gap of “affordable” for large travel websites and provides browsing directions for users who want to plan their own travel routes.