The Percentage Sampling transformation creates a sample data set by selecting a percentage of the transformation input rows. The sample data set is a random selection of rows from the transformation input, to make the resultant sample representative of the input.
The Percentage Sampling transformation is especially useful for data mining. By using this transformation, you can randomly divide a data set into two data sets: one for training the data mining model, and one for testing the model.
The Percentage Sampling transformation is also useful for creating sample data sets for package development. By applying the Percentage Sampling transformation to a data flow, you can uniformly reduce the size of the data set while preserving its data characteristics. The test package can then run more quickly because it uses a small, but representative, data set.
http://msdn.microsoft.com/en-us/library/ms139864.aspx
No comments:
Post a Comment