摘要
针对传统的旅游路线推荐算法推荐准确率与效率不高的问题,提出一种基于兴趣热点图的旅游路线推荐算法。通过离线挖掘出带滞留域与频繁路径的路径网络图,在此图基础上构建出融合时间约束因素的兴趣热点图,基于途中的旅程时间饱和度、景点密度比、景点热度比3个特征,提出候选子旅程排序TRR算法,选出候选全集K个得分最高的旅程路线作为结果用于推荐。仿真结果表明,相比PCF算法,TRR算法在推荐准确率与召回率方面更高,在算法执行效率方面更具优势。
To solve the problem of low precision and inefficiency in the traditional trip itinerary recommendation algorithm,a trip itinerary recommendation algorithm based on interesting hotspots graph(IHG)was proposed.The path map with the characte-ristics of stay points and frequent paths was mined,the IHG with time constraints was constructed on the basis of the map.Considering the three characteristics of travel time saturation(TTA),site density ratio(SDR)and sit heat ratio(SHR)in the journey,the rank algorithm TRR of the sub-trip set was proposed,the top-K trips was selected as a result for recommendation.The results of simulation show that the TRR algorithm is superior to the PCF algorithm in terms of recommendation accuracy and recall rate,and it has the advantages in the efficiency of algorithm implementation.
作者
陈健柯
陈平华
CHEN Jian-ke;CHEN Ping-hua(School of Computer,Guangdong University of Technology,Guangzhou 510006,China)
出处
《计算机工程与设计》
北大核心
2018年第9期2941-2946,共6页
Computer Engineering and Design
基金
广东省省级科技计划基金项目(2016B030308001
2016B030306002)
广州市科技计划基金项目(201604010099)
关键词
旅游路线推荐
兴趣热点图
位置定位轨迹
路径聚类
基于位置的服务
trip itinerary recommendation
interesting hotspots graph
GPS trajectories
path clustering
location-based service