With great innovation in technology there is a huge data explosion. Real-time surveillance, internet traffic, sensor data, health monitoring systems, communication networks, online transactions in the financial market and so on contribute as a data sources. Sometime data is huge enough that it cannot be stored in traditional databases. Moreover, this data can be structured, semi-structured or unstructured. In today’s scenarios it is always desired to take real time decisions from the data which is coming in with high velocity. Here, suitable solution is Stream processing. Stream processing allows us to analyze and mine data on-the-fly without storing it completely. The start point for the stream processing is the assumption that the potential value of data depends on data freshness. Thus, the stream processing paradigm analyzes data in real time to extract potential value out of it. Data mining techniques for streaming data includes: clustering, classification, frequent pattern mining and outlier detection which can be used to extract important information from streaming data.

In this paper we will summarize the efforts taken by researchers in the field of data stream mining along with the open research issues. We will also present the comparative study of few algorithms used for data stream processing. Finally we will conclude with the open issues in data stream processing.

©2016 International Journal of Computer Science and Applications 

Published by Research Publications, India