Detecting Anomalies for Large Sensors Data Contextually

Now-a-days processing and performing anomaly detection for big data is a very difficult and complex task. Large amount of data makes it difficult for earlier algorithms to obtain their real time characteristics. Most of the present  techniques to detect anomaly detection only consider the content of data source i.e. only concern about data itself instead of context of data .This paper describe a contextual anomaly detection framework which consist of two specific steps : content anomaly detection and context anomaly detection. Content detector is low priced computation technique and used to analyse anomalies in problem solving time. If content detector determines a sensor reading anomalous then context anomaly detection is used to trim the result of content detector. Context anomaly detection will be implemented by using profiles approach where profiles are class of identical data points achieved by applying multivariate clustering method.