具有信息补偿的移动用户群挖掘算法

Mobile user group mining algorithm with filling up missing data

针对移动网络中用户的移动、数据丢失、用户定位不精确导致的用户挖掘准确性低的问题,提出了一种基于卡尔曼滤波的噪声数据纠正与丢失数据补偿的移动用户群挖掘算法,提高空间与时间上具有相关性的移动用户群挖掘的有效性。详细的仿真实验以及与现有的基于距离的移动用户群挖掘算法DMUM的对比表明,该算法不仅具有更少的执行时间,同时有效地提高了移动用户群挖掘的召回率与准确率。

To overcome the low accuracy caused by mobility,missing data and imprecise location information,a mobile user group mining algorithm is proposed based on Kalman filter to correct noise data and fill up missing data.The proposed algorithm improves the effectiveness of spatial-temporal correlative user mining.Detailed simulation results and comparisons with the existed Distance-based Mobile User Mining（DMUM） algorithm show that the proposed algorithm is not only with less execution time,but also greatly improves recall and precision of mobile user mining.