Mining Structure Columns
Mining structure columns define how columns from a data source are used by a mining model in Microsoft SQL Server Analysis Services. Structure columns are added to the mining structure when it is created. Each mining structure column contains the following information:
A distribution is included if the column is continuous.
Modeling flags can also be defined on the mining model.
Third-party algorithms may also include custom properties that can be defined on the mining structure column.
Mining structure columns are designed to be flexible and extensible, because each algorithm that you use to build a mining model may use different columns from the structure to interpret the data.
The data types and content types that define structure columns are derived from the data source that you use to create the structure. You can change these settings within the mining structure, and you can also set modeling flags and set the distribution for continuous columns.
The following topics describe the different ways that you can define mining structure columns.
Describes the data types that you can use to define a mining structure column.
Describes the content types that you can use to define a mining structure column.
Describes the classified column types that you can use to relate one mining structure column to another.
Describes the distributions that you can use to define the data in a mining structure column.
Describes the different methods that you can use to discretize data in continuous mining structure columns.
Describes the modeling flags that you can use to define a mining structure column.
For more information about the data mining structure and the data mining model, see Mining Structures (Analysis Services - Data Mining). For more information about how algorithms are used to create mining models, see Data Mining Algorithms (Analysis Services - Data Mining).