Support Vector Regression

[#1]

 Build a boundary based statistical model to predict a continuous outcome as a function of multiple predictor variables. SVR is able to create much more flexible boundary shapes than the Regression Tree (Partition) method.

Support Vector Regression

1. From an open JMP^® table, select Analyze > Predictive Modeling > Support Vector Machines.
2. Add a continuous variable from Select Columns to the Y, Response role.
3. Add candidate predictor variables to the X, Factor role.
4. If desired, enter a validation column into the Validation role as shown in this example. Click OK.
5. The Model Launch control panel opens allowing a choice of a Kernal Function and associated options. Default settings were used for this example. Click Go. JMP displays:

• Model Summary detailing the performance of the model via metrics such as RASE and R-Squared. Here we see that the R-Squared is 89.6% - a rather good model at describing the outcome variable.
• Actual by Predicted Plot displaying the actual response values on the Y axis vs. the predicted response values on the X axis. This graph provides a graphical means to evaluate the predictive performance of the model, visualize the variation in the data around the fitted model, and identify outliers.

To visualize the model and the impact each factor has on the predicted response, launch the Profiler found under the red triangle. Contour and Surface Profilers are also available to visualization the relationship between pairs of the factors and the predicted response.

Additional options, such as Residuals Plots, Save Predicteds, Save Prediction Formula, and Publishing the Prediction Formula to the Formula Depot are accessible from the red triangle.

Diabetes.jmp (Help > Sample Data Folder)

Visit Predictive and Specialized Models > Support Vector Machines in JMP Help to learn more.