Peter Bartell peter.bartell@jmp.com suggests using Partial Least Squares when Ordinary Least Squares (Standard Least Squares in JMP) is not sufficient.
The PLS platform in JMP fits linear models based on linear combinations of the explanatory variables (Xs). These factors are obtained in a way that attempts to maximize the covariance between the Xs and the response or responses (Ys). PLS exploits the correlations between the Xs and the Ys to reveal underlying latent structures.
PLS is useful where the use of ordinary least squares does not produce satisfactory results including when you have more X variables than observations; highly correlated X variables; a large number of X variables; several Y variables; and many X variables.
JMP Pro provides additional functionality to conduct PLS Discriminant Analysis (PLS-DA), include a variety of model effects, use several validation methods, impute missing data, and obtain bootstrap estimates of the distributions of various statistics.
IN JMP Pro, PLS is a Fit Model personality that gives additional functionality over JMP by:
- fitting responses with a nominal and continuous data type
- fitting polynomial, interaction, and categorical effects
- offering a larger set of validation and cross validation methods including a train/validate/test construct, Kfold, Leave-One-Out and Holdback %
- imputing missing data
- providing Bootstrap estimates of distributions of select statistics.