Model Building

The process of building and scoring models by using database algorithms is similar to other types of data mining in IBM® SPSS® Modeler. The general process of working with nodes and modeling "nuggets" is similar to any other stream when working in IBM SPSS Modeler. The only difference is that the actual processing and model building is pushed back to the database.

A Database modelling stream is conceptually identical to other data streams in IBM SPSS Modeler; however, this stream performs all operations in a database, including, for example, model-building using the Microsoft Decision Tree node. When you run the stream, IBM SPSS Modeler instructs the database to build and store the resulting model, and details are downloaded to IBM SPSS Modeler. In-database execution is indicated by the use of purple-shaded nodes in the stream.