基于签到数据和交通工具的多目标旅游路线推荐

Multi-objective Trip Recommendation Based on Check-in Data and Traffic Tools

一般情况下,用户都希望花尽可能少的钱游览尽可能多的类别不同且得分高的景点,针对这一旅游路线推荐问题,提出了一种基于多约束多目标的剪枝(MCPA-MO)算法。首先,分析用户的历史签到数据,提出基于签到数据感知用户偏好和景点流行度的景点评分机制;然后,考虑在不同交通工具的选择下,景点之间会产生不同的转移时间和转移费用,进而影响时间和预算约束;最后,根据用户需求,使用MCPA-MO算法为用户计算出一条最优旅游路线。以Foursquare社交网站上的真实数据和马蜂窝上成都著名景点的信息作为数据集,实现并评估了推荐算法。实验结果表明,该方法能够为用户推荐更符合现实需求的个性化旅游路线。

In general, users want to spend little money to visit as many attractions with different categories and high scores as possible.In order to solve this problem, this paper proposes a multi-constrained and multi-objective pruning algorithm. First, it analyzes users1 historical check-in data, and proposes a rating mechanism of scenery spot based on users' check-in data-know users9 preferences and popularity of scenery spot. And this paper considers the different choices of vehicles, because the transfer between attractions will have different transfer time and costs. Finally, according to users9 needs, the MCPA-MO algorithm is used to calculate an optimal travel route for the user. The recommendation algorithm was implemented and evaluated using the real data on the Foursquare and the information on the famous attractions in Chengdu on the Horse Cell as the data set. Experimental results show that it can recommend personalized trip that meet the actual needs of users.