Share via

  • Article

Classified Columns (Data Mining)

You can define a classified column in a mining model in Microsoft SQL Server Analysis Services, to include information in that column that describes another column in the model. For example, you could define a classified column that contains standard deviations for another column in the model, such as for a column that describes a customer's total purchases for a calendar year.

The data type of a classified column must be either Long or Double.

The following list describes the content types that Analysis Services supports for classified columns.

  • PROBABILITY
    The value in the column is the probability of the associated value, and is a number between 0 and 1.

  • VARIANCE
    The value in the column is the variance of the associated value.

  • STDEV
    The value in the column is the standard deviation of the associated value.

  • PROBABILITY_VARIANCE
    The value in the column is the variance of the probability for the associated value.

  • PROBABILITY_STDEV
    The value in the column is the standard deviation of the probability for the associated value.

  • SUPPORT
    The value in the column is the weight, or case replication factor, of the associated value.

Note

The algorithms that Analysis Services provides do not support using classified columns.