Using GPS to learn significant locations and predict movement across multiple users
Personal and Ubiquitous Computing
Clustering of Time Series Subsequences is Meaningless: Implications for Previous and Future Research
ICDM '03 Proceedings of the Third IEEE International Conference on Data Mining
Discovering personally meaningful places: An interactive clustering approach
ACM Transactions on Information Systems (TOIS)
ACM Computing Surveys (CSUR)
Detection of Abnormal Crowd Distribution
GREENCOM-CPSCOM '10 Proceedings of the 2010 IEEE/ACM Int'l Conference on Green Computing and Communications & Int'l Conference on Cyber, Physical and Social Computing
Anomaly Detection on Collective Moving Patterns: A Hidden Markov Model Based Solution
ITHINGSCPSCOM '11 Proceedings of the 2011 International Conference on Internet of Things and 4th International Conference on Cyber, Physical and Social Computing
Performance comparisons of spatial data processing techniques for a large scale mobile phone dataset
Proceedings of the 3rd International Conference on Computing for Geospatial Research and Applications
Hi-index | 0.00 |
Anomaly detection is an important issue in various research fields. An uncommon trajectory or gathering of people in a specific area might correspond to a special event such as a festival, traffic accident or natural disaster. In this paper, we aim to develop a system for detecting such anomalous events in grid-based areas. A framework based on a hidden Markov model is proposed to construct a pattern of spatio-temporal movement of people in each grid during each time period. The numbers of GPS points and unique users in each grid were used as features and evaluated. We also introduced the use of local score to improve the accuracy of the event detection. In addition, we utilized Hadoop, a cloud-computing platform, to accelerate the processing speed and allow the handling of large-scale data. We evaluated the system using a dataset of GPS trajectories of 1.5 million individual mobile phone users accumulated over a one-year period, which constitutes approximately 9.2 billion records.