Hybrid Approach for Outlier Detection over Wireless Sensor Network Real Time Data

Outlier detection has gained considerable interest over real time data stream in data mining community with the realization that outliers can be the key discovery to be made from very large databases or data stream. Wireless sensor networks are one of active research area where massive amount of data measured or recorded. Those measurements that significantly deviate from the normal pattern of sensed data are considered as outlier's .The potential reasons of outliers include noise and errors, events, damage of device, and malicious attacks on the network. Sensor data is real time data recorded continuously with specific requirement and limitations of wireless sensor network in such condition traditional outlier detection techniques are not directly applicable. This paper provides an extensive review of existing outlier detection techniques specifically deployed for the wireless sensor networks and proposed a hybrid approach for detecting outlier in wireless sensor network. We proposed the technique to detect outlier over sensor data using a cluster based approach and distance based approach which contribute a compressive work for finding the outliers in real time data within WNS's.