predict.sparsenet {sparsenet}

make predictions from a "sparsenet" object.

Description

Similar to other predict methods, this functions predicts fitted values, coefficients and more from a fitted "sparsenet" object.

Usage

## S3 method for class 'sparsenet'
predict(object, newx, s = NULL,  which.gamma = NULL,
type=c("response","coefficients","nonzero"), exact = FALSE, ...)
## S3 method for class 'sparsenet'
coef(object,s=NULL, which.gamma = NULL,exact=FALSE, ...)

Arguments

object

Fitted "sparsenet" model object.

newx

Matrix of new values for x at which predictions are to be made. Must be a matrix. This argument is not used for type=c("coefficients","nonzero")

s

Value(s) of the penalty parameter lambda at which predictions are required. Default is the entire sequence used to create the model.

which.gamma

Index or indices of gamma values at which predictions are to be made. Default is all those used in the fit

type

"response" returns fitted predictions at newx. Type "coefficients" computes the coefficients at the requested values for s. Type "nonzero" returns lists of the indices of the nonzero coefficients for each value of s.

exact

By default (exact=FALSE) the predict function uses linear interpolation to make predictions for values of s that do not coincide with those used in the fitting algorithm. Currently exact=TRUE is not implemented, but prints an error message telling the user how to achieve the exact predictions. This is done my rerunning the algorithm with the desired values interspersed (in order) with the values used in the original fit

...

Not used. Other arguments to predict.

Details

The shape of the objects returned depends on which which.gamma has more than one element. If more than one element, a list of predictions is returned, one for each gamma.

Value

The object returned depends on type.

Author(s)

Rahul Mazumder, Jerome Friedman and Trevor Hastie

Maintainer: Trevor Hastie <[e-mail]>

References

Mazumder, Rahul, Friedman, Jerome and Hastie, Trevor (2011) SparseNet: Coordinate Descent with Nonconvex Penalties. JASA, Vol 106(495), 1125-38, https://hastie.su.domains/public/Papers/Sparsenet/Mazumder-SparseNetCoordinateDescent-2011.pdf

See Also

glmnet package, sparsenet, cv.sparsenet and print and plot methods for both.

Examples

x=matrix(rnorm(100*20),100,20)
y=rnorm(100)
fit=sparsenet(x,y)
predict(fit, which.gamma=5,type="nonzero")
predict(fit,x)