Real time big data mining

This thesis presents a parallel implementation of data streaming algorithms for multiple streams. Thousands of data streams are generated in different industries like finance, health, internet, telecommunication, etc. The main problem is to analyze all these streams in real time to find correlation between streams, standard deviation, moving average, etc. There are efficient algorithms available to analyze multiple streams. However, we can still improve the performance of a system to analyze multiple streams through parallel implementation. This thesis specifically focuses on: 1) design and implementation of a parallel system for multiple streams to find Discrete Fourier Transform (DFT), Most Correlated Pair, Singular Value Decomposition, Standard Deviation, Moving Average, and Aggregated Average; 2) performance analysis of multiple threaded application versus single threaded application; and 3) visualization of archived data.M.S.Includes bibliographical referencesby Jimit Pate