基于时空约束密度聚类的停留点识别方法

轨迹停留点的识别是轨迹分析、出行活动语义挖掘的关键。针对基于密度聚类的停留点识别方法对时空信息的表达缺陷，提出新的时空约束停留点识别方法，在密度聚类中引入轨迹的间接时空特征表示，将具有时空相似性的轨迹点进行聚合；采用与聚类过程相统一的时空特征约束对轨迹簇进行细粒度识别。算法在进行约束的时候再次利用到聚类时候所用的输入数据特征，特征的充分利用提高了识别的准确率。实验结果验证了本文方法的有效性。

Stay point recognition method based on spatio-temporal constraint density clustering

The recognition of the track stay point is the key to the trajectory analysis and the semantic mining of travel activities. Aiming at the defect of spatio-temporal information based on density clustering, the new method of space-time constrained stay point recognition is proposed. In the density clustering, the indirect spatio-temporal feature representation of the trajectory is introduced, and the trajectory points with spatio-temporal similarity are aggregated. The spatio-temporal feature constraint unified with the clustering process is used to fine grain the trajectory cluster. Therefore, when the constraints are used, the input data features used in the clustering are reused, and the full utilization of the features improves accuracy of the recognition. The experimental results verify effectiveness of the proposed method.