Properties for Mining Structure and Structure Columns

 

Applies To: SQL Server 2016

You can set or change the properties for a mining structure and for its associated columns and nested tables by using the Mining Structure tab of Data Mining Designer. Properties that you set in this tab are propagated to each mining model that is associated with the structure.

System_CAPS_ICON_note.jpg Note


If you change the value of any property in the mining structure, even metadata such as a name or description, the mining structure and its models must be reprocessed before you can view or query the model.

The following table describes the properties for the mining structure and the mining structure columns that are specific to data mining, and that you can view or configure in the Mining Structure tab. To view or configure these properties, right-click an element in the tree view, and then click Properties.

  • To view the properties of the structure, click the mining structure heading.

  • To view the properties of a column or a nested table, click the column name.

Properties of the Mining Structure

PropertyDescription
CacheModeSpecifies whether cases used in training should be cached or discarded after training is completed. Note: This property must be set to KeepTrainingCases to enable drillthrough and holdout.
CollationSpecifies the default collation for the column. If a collation is not specified, the collation of the server is used.
DescriptionDescribes the mining structure. As a best practice, the description should state the purpose and composition of the data in the structure.
ErrorConfiguration (default)Specifies options for special handling of errors, if any.
HoldoutMaxCasesSpecifies the maximum number of structure cases that can be reserved as a test data set. If values are specified for both HoldoutMaxCases and HoldoutPercent, the conditions are combined. Note: To set this property, CacheMode must be set to KeepTrainingCases.
HoldoutPercentSpecifies the percentage of the structure cases to reserve as a test data set. If values are specified for both HoldoutMaxCases and HoldoutPercent, the conditions are combined. Note: To set this property, CacheMode must be set to KeepTrainingCases.
HoldoutSeedSpecifies a seed to initialize partitioning of the holdout test set, to ensure that the test data set can be re-created. Note: To set this property, CacheMode must be set to KeepTrainingCases.
IDDisplays the unique identifier of the mining structure.

The name that you assigned to the mining structure when you created the structure is used as the ID. If you later change the name by typing a new value for the Name property, the new name is used as an alias only; the ID does not change.
LanguageSpecifies the language for the captions in the mining structure.
NameSpecifies the name or alias of the mining structure.

If you change the value for the Name property, the new name is used as a caption or alias only; the identifier for the mining structure does not change.
SourceDisplays the name of the data source, and the type of data source.

Properties of the Mining Structure Columns

PropertyDescription
ClassifiedColumnsIdentifies the column that a classified column describes.
ContentThe content type of the column.
DescriptionDescribes the column. As a best practice, the description of the column should provide information about how the data in the column has been derived or altered for data mining.
DiscretizationBucketCountDisplays the number of buckets in the discretized column.

Enabled only if the content type is set to Discretized.

This property is read-only.
DiscretizationMethodDisplays the method that was used to discretize the column.

Enabled only if the content type is set to Discretized.

This property is read-only.
DistributionSpecifies the distribution of content in the column.
IDDisplays the identifier of the column.

If you change the value of the Name property of the column, it does not affect the value of the ID property.
IsKeyIndicates whether the column is a key column.
KeyColumnsContains the definition of a column that is the key or is part of the key for an attribute.
ModelingFlagsSets additional parameters that are made available by the algorithm.
NameThe name of the column.
NameColumnIdentifies the column that provides the name of the parent element.
SourceDisplays the source of the column.

For relational data sources, the value is always (none).

For structures based on an OLAP cube, the value is the MDX statement that defines the slice used as the source for the nested table.
SourceMeasureGroupDisplays the source of the measure group.

For relational data sources, the value is always (none).

For structures based on an OLAP cube, the value is the MDX statement that defines the slice used as the source for the nested table.
TypeThe data type for the content in the column.

For more information about setting or changing properties, see Mining Structure Tasks and How-tos.

Create a Relational Mining Structure
Mining Structure Columns

Community Additions

ADD
Show: