计算机应用研究2017，Vol.34Issue(10)：2984-2988,5.DOI:10.3969/j.issn.1001-3695.2017.10.023
面向位置预测的动态轨迹模式挖掘
Dynamic trajectory pattern mining facing location prediction
摘要
Abstract
According to the research about a large amount of users' trajectory data,this paper proposed a method (PRED) to analysis trajectory pattern dynamically and predict location.The first step was using the improved pattern mining model to extract trajectory frequent patterns (named T-pattern).Then it put forward the DPTUpdate algorithm to design a data structure:DPT,which contained spatio-temporal information and could store and query the trajectory frequency pattern of moving objects.In addition,it presented the Prediction algorithm to calculate the optimal matching degree and got the predicted location of moving object trajectories.According to a series of contrast experiment based on the actual data,the PRED method can provide real-time analysis during the process,the average accuracy of this method reaches 72％ and the average coverage reaches 92.1％.The prediction effect has been increased significantly comparing with existing methods.关键词
轨迹模式/时空数据挖掘/模式树/位置预测Key words
trajectory pattern/spatio-temporal data mining/pattern tree/location prediction分类
信息技术与安全科学引用本文复制引用
邓佳,王永利,董振江..面向位置预测的动态轨迹模式挖掘[J].计算机应用研究,2017,34(10):2984-2988,5.基金项目
国家自然科学基金资助项目(61170035,61272420,61502233) （61170035,61272420,61502233）
国家科技重大专项资助项目(2012ZX03002003) （2012ZX03002003）
江苏省科技成果转化专项资金资助项目(BA2013047) （BA2013047）
江苏省六大人才高峰项目(WLW-004) （WLW-004）
兵科院预研项目 （）
中央高校基本科研业务费专项资金项目(30916011328) （30916011328）
国家级科研训练项目(201510288038) （201510288038）
