基于GPS轨迹的旅游兴趣点智能挖掘方法研究

Research on intelligent mining method of tourist points of interest based on GPS trajectory

为了给旅行者提供决策帮助,提出基于GPS轨迹的旅游兴趣点智能挖掘方法。采用GPS技术收集用户活动轨迹并聚类轨迹点,聚类点分类集合后进行分割,实现道路拟合,得到旅游兴趣点路网;通过用户和相似用户在此路网中的签到次数获取兴趣点,并归一化兴趣点评分结果,构建基于用户和社会关系的模型,解决协同过滤扩展问题,利用兴趣点流行度建立地理相关性模型进行二次过滤,求出用户可能感兴趣的兴趣点;采用标准变化加权法融合三种模型,构建兴趣点挖掘模型。实验结果表明,所提方法的内存分配合理,均方根误差低和AUC曲线低。

In order to provide decision-making help for travelers,an intelligent mining method of tourism interest points based on GPS trajectory is proposed.GPS technology is used to collect the user’s activity track and cluster the track points.After the cluster points are classified and collected,they are segmented to achieve road fitting and get the road network of tourism interest points.Through the number of check-in in the road network from users and similar users to obtain interest points,and normalize the score results of interest points to build a model based on the user and social relationship to solve the problem of collaborative filtering expansion.Meanwhile,the popularity of interest points is used to establish a geographical correlation model for secondary filtering,which could find out the interest points that users may be interested in.The standard change weighting method is used to fuse the three models to construct the interest point mining model.Experiment results show that the proposed method has reasonable memory allocation,low root mean square error and low AUC curve.