Abstract: A content-based approach is proposed in this paper in order to meet the growing demand for exercise route recommendation in mobile exercise management applications. It considers not only the similarities among the routes but also those of the users’ abilities in order to make personalized route recommendation. In addition, it provides a solution to the cold-start problem when applied in practice. Experimental results show that this approach is capable of providing users with good personalized recommendation of exercise routes with high efficiency.

Key words: exercise route recommendation, content similarity, cold start, Geohash

摘要： 运动管理类移动应用经常需要从大量运动轨迹中为用户搜索出适合的运动路线，为此，提出了基于内容相似度的运动路线推荐方法。利用Geohash编码快速搜索用户邻域范围内的所有待推荐路线，计算用户已有运动路线与待推荐路线之间的路线相似度和运动模式相似度，对两种相似度进行加权求和处理作为待推荐路线的相似度，将相似度较大的若干运动路线推荐给用户。还针对推荐存在的冷启动问题给出了解决方案。通过实验，对两种相似度的权值进行调优，实验结果显示该方法可提高推荐效率，实际推荐效果良好。

关键词: 运动路线推荐, 内容相似度, 冷启动, Geohash