predict.npmr {npmr}

Make predictions from a “npmr” object

Description

Return predicted reponse class probabilities from a fitted NPMR model, for each value of lambda on which the NPMR model was originally fit.

Usage

## S3 method for class 'npmr'
predict(object, newx, ...)

Arguments

object

an object of class npmr

newx

covariate matrix on which for which to make response class probability predictions. Must have same number of columns as X used original to fit object.

...

ignored

Value

a 3-dimensional array, with dimensions (nrow(newx), ncol(Y), length(lambda)). For each lambda, this array stores for that value of lambda the predicted response class probabilites for each observation.

Author(s)

Scott Powers, Trevor Hastie, Rob Tibshirani

References

Scott Powers, Trevor Hastie and Rob Tibshirani (2016). “Nuclear penalized multinomial regression with an application to predicting at bat outcomes in baseball.” In prep.

See Also

npmr, predict.cv.npmr

Examples

#   Fit NPMR to simulated data

K = 5
n = 1000
m = 10000
p = 10
r = 2

# Simulated training data
set.seed(8369)
A = matrix(rnorm(p*r), p, r)
C = matrix(rnorm(K*r), K, r)
B = tcrossprod(A, C)            # low-rank coefficient matrix
X = matrix(rnorm(n*p), n, p)    # covariate matrix with iid Gaussian entries
eta = X 
P = exp(eta)/rowSums(exp(eta))
Y = t(apply(P, 1, rmultinom, n = 1, size = 1))

# Simulate test data
Xtest = matrix(rnorm(m*p), m, p)
etatest = Xtest 
Ptest = exp(etatest)/rowSums(exp(etatest))
Ytest = t(apply(Ptest, 1, rmultinom, n = 1, size = 1))

# Fit NPMR for a sequence of lambda values without CV:
fit2 = npmr(X, Y, lambda = exp(seq(7, -2)))

# Compute mean test error using the predict function (for each value of lambda):
getloss = function(pred, Y) {
    -mean(log(rowSums(Y*pred)))
}
apply(predict(fit2, Xtest), 3, getloss, Ytest)