The Mining Wizard enables you to complete the following actions:
Create or modify a build task—A build task specifies all information necessary to train an algorithm against a set of data.
Create or modify a test task—A test task applies a data mining model to a set of data that correlates with known results. Comparing the results generated by the model to the known results provides information on how accurately the model is performing.
Create or modify an apply task—An apply task applies a data mining model to a new set of data to extract new information. The apply task writes the results back to the cube.
Create or modify a scoring task—A scoring task applies a data mining model to a new set of data to extract new information. The results are not written back to the cube. Input to a scoring task can be preset in a cube or you can enter the input when executing the template.
When you create a task, you can save the task settings in a named template which you can reuse or modify for later tasks.
You can run this wizard in Advanced or Simplified mode. Advanced mode enables you to apply transformations to accessors and to map them. Fewer accessor values are required in Simplified mode.
Note: | Data mining is not supported for aggregate storage databases. |
The Mining Wizard steps you through the process of creating, testing, and applying a model to extract information from your database. This topic contains the following sections:
“MDX” in the Oracle Essbase Technical Reference