Select more predictors for partial dependence plots for Discover Best Model (Continuous Response)

Run Predictive Analytics Module > Automated Machine Learning > Discover Best Model (Continuous Response). Select One-Predictor Plots or Two-Predictor Plots in the results.

Note

This command is available with the Predictive Analytics Module. Click here for more information about how to activate the module.

If the best model type is a TreeNet^® model, then the results include partial dependence plots. If partial dependence plots are available from the model that are not in the results, then you can select One-Predictor Plots or Two-Predictor Plots in the results. Available plots can be plots that were not in the original results or deleted plots. The buttons are active only when the worksheet for the model is in the project.

Add one predictor partial dependence plots

You can add a plot for any single predictor in the model that is not already on a plot.

Select One-Predictor Plots in the results.
In Predictors, enter the predictors to plot.
Click OK.

The new plots are at the end of the section for One-Predictor Partial Dependence Plots.

Add two predictor partial dependence plots

You can add a plot for any pair of predictors in the model that is not already on a plot. Initially, the table of predictors includes the 5 strongest interactions in the model that do not have a plot in the results. Complete the following steps to add plots:

Select Two-Predictor Plots in the results.
In a row of the table, enter a predictor for the x-axis of the plot in X Predictor. Then, enter a predictor in the same row for the y-axis of the plot in Y Predictor.
If any row in the table includes a pair of two continuous predictors, then the choices in Types activate. Select whether to create surface plots, contour plots, or both.
Click OK.

The new plots are at the end of the section for Two-Predictor Partial Dependence Plots.