DataMining Digest

Digest Of DataMining With SQLServer 2005

1.        There are two kinds of data mining techniques: supervised and unsupervised. Supervised data mining requires the user to specify a target attribute and a set of input attributes. The typical supervised data mining algorithms include decision trees, Naïve Bayes, and neural networks. An unsupervised data mining technique doesn’t have to have a predictable attribute. Clustering is a good example of unsupervised data mining. It groups heterogeneous data points into subgroups so that data points in each subgroup are more or less homogeneous.

 

2.        The OLAP mining model often contains nested tables. The case table of an OLAP mining model is always one of the dimensions and nested tables always come from one of the fact tables using a another dimension attribute as the nested key.

 

3.        Mining models defined based on a relational source and a multidimensional source have the same structure and metadata. The only difference is the way the models are bound to data and processed. In fact, a model created and processed with relational tables can be reprocessed with new bindings to a cube.

4.        SSIS Transforms

5.        Tasks and Transforms for Data Mining