predict.smvcir: Predict new data using SMVCIR
In danno11/SMVCIR: Sliced Mean Variance Covariance Inverse Regression

Takes a set of observations and projects them into the SMVCIR discriminant space based on a previously built SMVCIR model object. Probabilities are calculated by measuring the mahalonobis distance of an observation from each of the class centroids based on the original SMVCIR model object.

## S3 method for class 'smvcir'
predict(model, newdata, type = "prob",
  coordinates_only = FALSE, method = "centroid",
  maxdim = model$predictors, kernel = NULL)

`type`	Either "prob" for class probabilities, or "class" for class predictions.
`coordinates_only`	Default is FALSE. If TRUE, SMVCIR discriminant coordinates are returned and can be used towards other classification techniques, such as SVM's or logistic regression.
`maxdim`	Desired number of smvcir dimensions. A dimensionality test from previously built model may indicate a sufficient number of dimensions, by default, all possible dimensions are included.