A Survey on Outlier Detection Techniques in Dynamic Data Stream

Outlier detection has significant importance in the data mining domain. Applications which contain streaming data flow may have many abnormal or outlier data and these applications require efficient outlier detection techniques to detect and analyze these abnormal patterns. Outlier detection is the process of detecting patterns in the data which do not adhere to the normal behavior or data. These patterns are known by several terms such as anomalies, outliers, noise or inconsistent data. Detecting and analyzing the abnormal data like outliers is a wide research area with tremendous applications. Finding and selecting appropriate detection technique is mandatory. This survey presents the tools and techniques used for detecting outliers in data streams and attempts to classify the problem in outlier detection methods over the data stream. The review of detection techniques gives an insight into the further research opportunities in this domain.