Caching for Multi-dimensional Data Mining Queries (2001)
Abstract:
Multi-dimensional data analysis and online analytical processing are standard querying techniques applied on today’s data warehouses. Data mining algorithms, on the other hand, are still mostly run in stand-alone, batch mode on flat files extracted from relational databases. In this paper we propose a general querying model combining the power of relational databases, SQL, multidimensional querying and data mining. Using this model allows data mining to leverage much of the extensive infrastructure that has already been built for data warehouses including many of the highly successful query processing strategies designed for OLAP. We present one such integrated, chunk-based caching scheme that is central to the design of an interactive, multi-dimensional data mining system and conclude with an experimental evaluation of three different cache replacement algorithms. 1
Citations
| 1607 | Fast Algorithms for Mining Association Rules – Agrawal, Srikant - 1994 |
| 121 | Psaila G., A New SQL-like Operator for Mining Association Rules – Ceri, Meo - 1996 |
| 30 | Using a knowledge cache for interactive discovery of association rules – Nag, Deshpande, et al. - 1999 |
| 23 | Integrating Mining with Relational Database Systems: Alternatives and Implications – Sarawagi, Thomas, et al. - 1998 |
| 11 | Exploratory Mining and Pruning – Ng, Lakshmanan, et al. - 1998 |
| 4 | A Framework for Measuring Changes – Ganti, Gehrke, et al. - 1999 |
