Tolerance-Based Adaptive Online Outlier Detection for Internet of Things

The cheap and low-quality sensor devices are usually used for event detection in Internet of Things (IoT), and they put limitations on power, memories and computing capabilities. Those limitations need to be considered while designing our outlier detection algorithm. In this paper, we try to present an adaptive online outlier detection algorithm to handle data measurements for event detection. Namely, the proposed algorithm needs to provide the capability to tolerant those data which would be classified as outliers by traditional algorithms. With an accurate ratio of outliers and a tolerance parameter, a tolerance-based adaptive online outlier detection (TAOOD) algorithm is proposed. The contributions of TAOOD are two folds: (i) TAOOD decreases the amount of transmitted data by discarding duplicate data and outliers, (ii) TAOOD eliminates the limitation of original window-based outlier detection algorithm by adapting an accurate ratio of outliers and a tolerance parameter. Extensive simulations demonstrate effectiveness of the proposed algorithm.