JMP 14.0 Online Documentation (English)
Discovering JMP
Using JMP
Basic Analysis
Essential Graphing
Profilers
Design of Experiments Guide
Fitting Linear Models
Predictive and Specialized Modeling
Multivariate Methods
Quality and Process Methods
Reliability and Survival Methods
Consumer Research
Scripting Guide
JSL Syntax Reference
JMP iPad Help
JMP Interactive HTML
Capabilities Index
JMP 13 Online Documentation
JMP 12 Online Documentation
Multivariate Methods
•
Discriminant Analysis
•
Discriminant Launch Window
• Shrink Covariances
Previous
•
Next
Shrink Covariances
In the Discriminant launch window, you can select the option to Shrink Covariances. This option is recommended when some groups have a small number of observations. Discriminant analysis requires inversion of the covariance matrices. Shrinking off-diagonal entries improves their stability and reduces prediction variance. The Shrink Covariances option shrinks the off-diagonal entries by a factor that is determined using the method described in Schafer and Strimmer (
2005
).
If you select the Shrink Covariances option with the Linear discriminant method in the launch window, this provides a shrinkage of the covariance matrices that is equivalent to the shrinkage provided by the Regularized discriminant method with appropriate Lambda and Gamma values. When you select the Shrink Covariances option and run your analysis, the Shrinkage report gives you an Overall Shrinkage value and an Overall Lambda value. To obtain the same analysis using the Regularized method, enter 1 as Lambda and the Overall Lambda from the Shrinkage report as Gamma in the Regularization Parameters window.
Previous
•
Next
Help created on 7/12/2018