MSMiner is a multi-strategy data mining platform based on data warehouse. Here, the data of different versions, granularities in warehouse, which is managed by meta-data, can be updated and mined step by step. Over ten mining algorithms, such as AQ, Decision Tree, NN, RS, Statistics, SVM, Association Rule, Bayesian theory, and Visualization, can be used in the most mining tasks, i.e. prediction, classification, clustering, AR, series patterns, dependence relation, and fraud detection. Wizard guiding schema manages all verbose details and facilitates users' interactive and explorative operation. Moreover, mining algorithms are encapsulated, modularized, and extensible. It is convenient, flexible, and effective to combine it with real domain and build an application system, in which verification knowledge and exploration knowledge can be discovered effectually.
KDD is the
iterative, interactive, nontrivial
process of identifying
valid, novel, potentially useful, and ultimately
understandable
patterns
in data.
