Data mining architectures - A comparative study

Data mining is the process of deriving knowledge from data. The architecture of a data mining system plays a significant role in the efficiency with which data is mined. It is probably as important as the algorithms used for the mining process. CRITIKAL is a three-tier data mining architecture consisting of Client, Middle tier and the Data Warehouse. The architecture for mining semi-structured data makes a distinction between structured and unstructured data, and uses separate storage areas for them. The Kensington data mining system is an internet-based mining system for the analysis of large and distributed data sets. The Matheus et al.'s Multicomponent Architecture is designed to perform spatial data mining while DARWIN and PaDDMAS are used for distributed data mining. Lastly the architecture for scientific data mining, is used to mine scientific data from large science archives. The main factors that have been focused on, in these architectures are portability, scalability reduction in data preparation time, integration, multi-strategy and distribution.