Data mining — Defining mining tasks

A mining task brings together the information that is required to start a training run and to compute the mining model. This information consists of the mining settings and the input data definition.

Intelligent Miner® provides user-defined methods to define mining tasks.

You can also define tasks for test runs. These tasks bring together all the information actually needed to start a test run. This information consists of the mining data used to test a mining model and a location of the mining model to be tested.

The following sections provide SQL code examples for the definition of the individual mining tasks:

Depending on the mining task you are defining, one of the following data types is involved:

DM_RuleBldTask
DM_ClasBldTask
DM_ClusBldTask
DM_RegBldTask
DM_TsBldTask
DM_ClasTestTask
DM_RegTestTask

A DM_RuleBldTask, DM_ClasBldTask, DM_ClusBldTask, DM_RegBldTask, or DM_TsBldTask type provides all the information actually needed to start a training run and to compute a data mining model. In a task, a settings definition is combined with a physical data specification of type DM_MiningData. This combination is valid if an alias name is defined in the physical data specification for each field name in the logical data specification of the settings values.