Data Mining Techniques in Parallel and Distributed Environment- A Comprehensive Survey

Distributed sources of voluminous data have raised the need of distributed data mining. Conventional data mining techniques works well on structured data which is clean, pre-processed and properly arranged either in the form of structured files, databases or data warehouse. These techniques are based upon centralised data store however they have several limitations in distributed scenario where the data is scattered in different geographical locations on data servers all across the network. It becomes a costly affair to accumulate huge data on a centralised node in real time. To overcome these limitations, application of distributed data mining techniques has become essential. This paper describes various data mining tools and techniques that can be used in distributed environment. Different algorithmic and architectural approaches are followed in various distributed mining techniques. Latest approaches in distributed data mining are explored. Various research issues and challenges in the field of distributed data mining are also discussed. Abbreviations: KDD-Knowledge discovery in databases, ARMAssociation rule mining, DDMDistributed Data Mining, GPU-Graphical processing Unit