Data mining is available only when DB2 Intelligent Miner is installed.
You can create and register data mining models.
Creating data mining models
You can create and modify data mining models (also known as settings objects) by using DB2 Intelligent Miner for Data.
Data mining models have built-in relationships to other objects in the same mining base. For example, a mining settings object requires that you specify the input data, either by creating a new data object or by referring to an existing one. The mining settings object represents parameters that you specify for a mining function. One of these parameters is the name of the data object, which serves as a logical description of the input data.
Any user-defined input or output table uses the default data mining table schema wcamng. Any user-defined input or output table for data mining should use the prefix UX_ (for user extension). If you define the setting name for the model training as ux_xxxx, the corresponding setting for applying this model will be named ux_xxxxa.
To create new data mining models, refer to the IBM DB2 Intelligent Miner for Data V8.1 Using the Intelligent Miner for Data documentation. Also refer to Selecting mining models for more information.
Registering new data mining models
Data that is used in mining is identified and grouped into a set of tables or views. Use these tables and views as the input data for mining.
When you add a new data model, register it in the following control tables:
- DATATABS: Tables for both input and output are modeled in the data control table. Register each data set with one entry in this table.
- MODELS
- APPTABS