Testing and Validation Tasks and How-tos (Data Mining)

Applies to: SQL Server 2019 and earlier Analysis Services Azure Analysis Services Fabric/Power BI Premium

You can use the Mining Accuracy Chart tab of Data Mining Designer in SQL Server Data Tools to compare the predictive accuracy of the mining models in your mining structure.

You can create four kinds of charts:

• Lift chart

• Profit chart

• Classification matrix

• Cross-validation report

The first three charts use the Input Selection tab to define the data that is used for generating the chart.

The Cross-validation chart is created by using additional inputs, available on the Cross-Validation tab. For more information, see Cross-Validation (Analysis Services - Data Mining).

For more information about how to use the mining accuracy chart, see Testing and Validation (Data Mining).

In This Section

• Create a Lift Chart, Profit Chart, or Classification Matrix

• Create a Cross-Validation Report

• Choose and Map Model Testing Data

• Apply Filters to Model Testing Data

• Choose the Column to Use for Testing a Mining Model

• Using Nested Table Data as an Input for an Accuracy Chart