A survey of outlier detection algorithms for data streams
Jinita Tamboli, Madhu Shukla · International Conference on Computing for Sustainable Global Development · 2016
Data mining is characterized as the process of examining hidden patterns and outlining into some useful information. It is an exciting field of research for researchers. Data streams are continuous instance of records and mining interesting knowledge from this instance is known as data stream mining. Outlier detection is currently an important research problem in many fields and is also involved in many of the applications. Outlier detection in streaming data is a challenging task as only one scan is possible and they need huge amount of storage which is practically infeasible. There are many existing methods for outlier detection based on distance measure but are not efficient for data stream as they are dynamic in nature. This paper discusses on various algorithms for outlier detection on data streams.