GM（1，1）模型在轨迹聚类中的应用

针对移动对象轨迹数据在获取过程中可能存在延迟、缺失,使得轨迹数据存在不确定性的情况.利用GM(1,1)模型在预测、决策领域中的优势,在拟合误差阀值的限制下,提出一种基于GM(1,1)模型的轨迹分割方法(TR_GMPR).之后,对分割后的轨迹段利用DBSCAN算法进行聚类.实验表明,生成的特征轨迹相比其他线段分割的轨迹聚类结果,更符合实际情况.

Application of model GM(1,1) in the trajectory clustering

Moving objects trajectory data may become uncertainty because of being lost or delayed in transmission process.In this paper, using the superiority of Model GM(1,1) in domain of forecasting or decision-making field , a new method of trajectory segmentation based on Model GM(1,1)(TR_GMPR) was proposed within the limitations of fitting error threshold. Then, the DBSCAN algorithm was used to cluster the segmented trajectories, which produced the characteristic trajectories. The experiment indicated that these characteristic trajectories could be further accord with the practical situation ,ompared to other clustering results adopted the line segmentation method .