CAJ | 학술논문

在位置服务领域，用户轨迹在较大程度上体现了用户的日常行为模式，以及个人生活习惯等。利用GPS终端收集用户行为轨迹数据并加以挖掘分析，对于位置服务实现智能化推送有积极作用。用户行为轨迹的停留点分析是轨迹分析的常见手段之一。本研究首先将用户个性化信息，与轨迹点相关的地标名称等语义信息融入常规用户行为轨迹，形成“位置-语义”一体化的用户语义轨迹。然后，过滤原始轨迹错误点，提高数据精度，并在此基础上采用一种新的加权方法计算轨迹停留点坐标。最后，利用停留点坐标结合用户的兴趣、职业等个人信息，在扩充的POI信息库（包含营业时间、优惠信息等）中检索匹配，并智能化匹配出用户停留点周围的POI，主动向用户推送符合个人兴趣或职业需求的POI详情位置服务。
재위치복무영역，용호궤적재교대정도상체현료용호적일상행위모식，이급개인생활습관등。이용GPS종단수집용호행위궤적수거병가이알굴분석，대우위치복무실현지능화추송유적겁작용。용호행위궤적적정류점분석시궤적분석적상견수단지일。본연구수선장용호개성화신식，여궤적점상관적지표명칭등어의신식융입상규용호행위궤적，형성“위치-어의”일체화적용호어의궤적。연후，과려원시궤적착오점，제고수거정도，병재차기출상채용일충신적가권방법계산궤적정류점좌표。최후，이용정류점좌표결합용호적흥취、직업등개인신식，재확충적POI신식고（포함영업시간、우혜신식등）중검색필배，병지능화필배출용호정류점주위적POI，주동향용호추송부합개인흥취혹직업수구적POI상정위치복무。
Location Based Service (LBS), with the support of GIS, is a thriving service for users related with co-ordinates received by wireless communication network or GPS. The trajectories composed with a set of received coordinates mainly express the character and habit of user’s behavior. Through analyzing and mining users’tra-jectories, we will improve the efficiency of location based service. In this paper, firstly, the trajectories data in-cluding location coordinates and semantic fields are collected through GPS signal by the self-developed software installed in the terminal. The semantic fields contain the ID of user, current speed, nearby landmark and so on. Then the mistakes incorporated in raw trajectories due to the GPS instability should be filtered to enhance data accuracy. A method has been applied to filter the“jitter”points and to calculate the angle (angle threshold is 15°) and time interval (time threshold is 3s). Different from the conventional method that calculates mean value as the stay point’s coordinate directly, we divide the points in sub-trajectory into different groups based on semantic in-formation. Afterwards, on the basis of the number of points in each group, we acquire weighted coordinate of the stay point. Finally, we match the stay points with POIs, which have ample information, like opening hours, spe-cial offers, etc., and then get a set of matched POIs around the stay point. In addition, through analyzing the inter-est and job of user, it could retrieve the more appropriate service and send it to user accordingly.