Booking-com/multi-destination-trip-dataset

Booking.com提供了一个基于数百万真实匿名预订的独特数据集，旨在鼓励对序列推荐问题的研究。训练数据集包含1,166,835条匿名酒店预订记录，每条记录包括用户ID、入住日期、退房日期、创建日期、联盟渠道ID、设备类型、预订国家、酒店国家、城市ID和旅行ID等特征。评估数据集类似构建，但每个旅行的最后一个预订的城市ID和国家被隐藏，需要预测。

Booking.com provides a unique dataset based on millions of real anonymized bookings to encourage the research on sequential recommendation problems. The dataset consists of over 1,166,835 anonymized hotel reservations, each with features such as user_id, check-in date, device class, bookers country, hotel country, city_id, etc. Each reservation is part of a users trip (identified by utrip_id), which may include multiple consecutive bookings. The evaluation dataset is structured similarly but conceals the city_id of the final reservation of each trip, requiring prediction.