Outlier detection: A survey on techniques of WSNs involving event and error based outliers

Abstract

In the recent few years, many wireless sensor networks have been distributed systematically in the real world to collect valuable raw sensed data. However, the crucial point of challenge is to extract high level knowledge from this raw sensed data. In the application of data analysis, a necessary preprocessing step is anomaly detection, also known as deviation detection or data cleansing. Outliers in wireless sensor networks (WSNs) are those measures that deviate from a defined pattern. Outlier detection can be used to remove noisy data, detect faulty nodes and discover interesting events. Numerous small and low cost nodes loaded with capabilities of integrated sensing and computation are involved in a WSN structure. Due to high density WSNs are exposed to faults and nasty attacks causing inaccurate and unreliable sensors reading, making Wireless sensor networks prone to outliers. This survey provides an outline of outlier detection techniques and approaches focusing on event and error based outliers.