Big Trajectory Data Analysis for Clustering and Anomaly Detection
Hirokatsu Kataoka, Yoshimitsu Aoki, Kenji Iwata, Yutaka Satoh, Ikushi Yoda, Masaki Onishi · 2013
We’ve been developing a sensor that can acquire positional data. Recently, a position-based big data creation is easy task and trajectory analysis is the highest priority for ”position-based service”. Traffic congestion, marketing mining, and pattern analysis are the one of the examples in trajectory analysis field. In this paper, we propose the trajectory analysis approach for clustering and anomaly detection by using big trajectory data. To execute clustering, we understand an environment in front of the camera and set a cluster route from trajectory map. The experiment shows that the proposed method understands environment and performs clustering. Moreover, the approach classifies anomalies from big data.