predict.pliable: Compute predicted values from a fitted pliable object Make predictions from a fitted pliable lasso model

Description

Compute predicted values from a fitted pliable object Make predictions from a fitted pliable lasso model

Usage

# S3 method for pliable
predict(object, x, z, type = c("link", "response",
  "coefficients"), lambda = NULL, verbose = FALSE, ...)

Arguments

object

object returned from a call to pliable

n by p matrix of predictors

n by nz matrix of modifying variables. These may be observed or the predictions from a supervised learning algorithm that predicts z from test features x and possibly other features. See example below

type

Returns either the fitted values with type="link" or "response" or the parameter estimates when type="coefficients" . Type "link" gives the linear predictors for binomial, or cox models; for gaussian models it gives the fitted values. Type "response" gives the fitted probabilities for binomial, and the fitted relative-risk for cox; for gaussian type "response" is equivalent to type "link".

lambda

values of lambda at whcih predictions are desired. If NULL (default), the path of lambda values from the fitted model. are used. If lambda is not NULL, the predictions are made at the closest values to lambda in the lambda path from the fitted model;

verbose

Should information should be printed along the way? Default FALSE.

...

Further arguments (not used) @return predicted values

Examples

Run this code

# NOT RUN {
# Train a pliable lasso model
n = 20; p = 3 ;nz=3
x = matrix(rnorm(n*p), n, p)
z =matrix(rnorm(n*nz),n,nz)
y = x[,1] +x[,1]*z[,3]+ rnorm(n)
  fit = pliable(x,z,y)
 # plot coefficient profiles, indicating z-interactions with a "x" symbol
plot(fit)


# Predict using the fitted model
ntest=500
xtest = matrix(rnorm(ntest*p),ntest,p)
ztest =matrix(rnorm(ntest*nz),ntest,nz)

pred= predict(fit,xtest,ztest)

 #Example where z is not observed in the test set,  but predicted from a  supervised
  #  learning method

 library(glmnet)
n = 20; p = 3 ;nz=3
x = matrix(rnorm(n*p), n, p)
z =matrix(rnorm(n*nz),n,nz)
y = x[,1] +x[,1]*z[,3]+ rnorm(n)
fit = pliable(x,z,y)
# predict z  from x; here we use glmnet, but any other supervised learning method
# could be used
zfit=glmnet(x,z,family="mgaussian")

# Predict using the fitted model
ntest=500
xtest = matrix(rnorm(ntest*p),ntest,p)
ztest =predict(zfit,xtest,s=zfit$lambda.min)[,,20]

pred= predict(fit,xtest,ztest)

# }

Run the code above in your browser using DataCamp Workspace