Density Based Cluster Technique for Sensor Data

Abstract

Knowledge discovery from sensor data is becoming increasingly important in many applications such as national security, environment, energy, smart homes, and so on. Sensor devices generates large amount of spatial and non-spatial data. Due to the characteristics of sensor networks, existing data mining methods cannot be applied. In this paper we present the enhancement of GDBSCAN for mining the sensor data. This method considers both spatial and non-spatial information together. Also the efficiency of algorithm is measured. Experimental evaluation is presented on synthetic and real data to demonstrate the effectiveness and robustness of the proposed algorithm.

Keywords: Sensor Data; KDD-Sensor; Data Mining; Clustering; Density Based Cluster.