Parallel Association Rule Mining System for Continuous Data Streams

In recent times, the data generation and storage modules in several organizations have seen a shift from the centralized approach to a parallel and distributed architecture. Also, the nature of data is changing from the static tables in databases and data warehouses to continuous stream data in data stores used for various technical and business applications. In such times, it is necessary to have data mining systems which accept data generated independently at disparate sites and perform mining at the local sites as well as at the global level at a centralized location. This paper describes a complete system to handle continuous stream data at parallel sites, the memory management involved in the storage of this data and an algorithm specifically designed to perform association rule mining in such an environment.