基于频繁序列挖掘的后续行程序列推荐  

Successive Travel Sequence Recommendation Based on Frequent Sequence Mining

作　　者：温彦[1] 马立健 陈明 WEN Yan;MA Li-jian;CHEN Ming(College of Computer Science and Engineering,Shandong University of Science and Technology,Qingdao 266590,China;State Grid Shandong Electric Power Company,Qingdao Power Supply Company,Qingdao 266500,China)

机构地区：[1]山东科技大学计算机科学与技术学院 [2]国网山东省电力公司青岛供电公司

出　　处：《软件导刊》2019年第3期53-56,共4页Software Guide

基　　金：教育部人文社会科学研究青年基金项目(17YJCZH187);国家自然科学基金项目(61702306;61602278);青岛市哲学社会科学规划项目(QDSKL1801131)

摘　　要：个性化旅游发展迅速,已有方法主要集中在单个旅游产品推荐上,而旅游行程存在明显的序列性,并受到当前已有行程轨迹影响。因此,提出一种旅行中后续行程序列的推荐方法 SeqRem,基于所有用户的行程序列挖掘频繁序列模式,并以此为依据利用最大点权独立集方法对用户的历史行程序列进行分割,以发现最优序列推荐内容。实验证明,SeqRem在单点推荐和序列推荐准确率与召回率均具有较好效果。Personalized tourism has become the tendency.Current approaches mainly focus on single product recommendations.However,the travel routes have obvious sequential characteristics,and they are affected by existing routes.A successive travel sequence recommendation approach SeqRem is proposed,it mines the frequent sequence pattern of all the travelers,and use maximum weighted in.dependent set to cut a user’s historical routes while discover optimum successive sequence.Experiments demonstrate that SeqRem performs well in terms of precision and recall rates.

关 键 词：推荐系统 频繁序列挖掘 兴趣点 后续行程序列 数据挖掘 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]