研究个体在不同时间的行为规律性,以及不同个体行为之间的相似性,可以为个性化推荐以及基于位置的服务提供帮助.从手机的基站位置数据中,通过聚类方法找到参考位置,并根据参考位置,将人们杂乱无章的行为转变为到达和离开的二进制时间序列.定义二进制时间序列的相似度,利用异或算法检测个体行为模式.在Reality数据集上的实验结果表明,该方法是有效且可靠的.
The regularity of the behavior of the same individual at different times and the similarity of different individual behaviors can provide help for personalized recommendation and loca- tion-based services. According to the location data of the mobile phone, the reference position is found by the clustering method. And then people's behavior is transformed into the arrival and de- parture of the binary time series based on the reference position. The similarity of binary sequences is defined and then individual behavior patterns are detected using XOR algorithm. Experiments on Reality mining data sets show that the proposed method is effective and reliable.