Output details

Mining Recurring Concepts in a Dynamic Feature Space

The paper makes an important contribution to the data mining research by dynamically changing the feature space according to the user’s context. The thorough experimental study shows that the proposed technique is the fastest to recover when changes in the underlying concept generating the streaming data recurs, setting a benchmark in this research area. Parallelisation and high performance computing approaches have been used to address the Big Data problems; however, the technique provided in this research provides a novel and significant approach to Big Data analytics.