CRAN Package Check Results for Package personalized

Last updated on 2018-02-25 00:48:53 CET.

Check Details

Version: 0.1.3
Check: examples
Result: ERROR
    Running examples in ‘personalized-Ex.R’ failed
    The error most likely occurred in:
    
    > ### Name: check.overlap
    > ### Title: Check propensity score overlap
    > ### Aliases: check.overlap
    >
    > ### ** Examples
    >
    > library(personalized)
    >
    > set.seed(123)
    > n.obs <- 250
    > n.vars <- 15
    > x <- matrix(rnorm(n.obs * n.vars, sd = 3), n.obs, n.vars)
    >
    >
    > # simulate non-randomized treatment
    > xbetat <- 0.25 + 0.5 * x[,11] - 0.5 * x[,12]
    > trt.prob <- exp(xbetat) / (1 + exp(xbetat))
    > trt01 <- rbinom(n.obs, 1, prob = trt.prob)
    >
    > # create function for fitting propensity score model
    > prop.func <- function(x, trt)
    + {
    + # fit propensity score model
    + propens.model <- cv.glmnet(y = trt,
    + x = x, family = "binomial")
    + pi.x <- predict(propens.model, s = "lambda.min",
    + newx = x, type = "response")[,1]
    + pi.x
    + }
    >
    > check.overlap(x = x,
    + trt = trt01,
    + propensity.func = prop.func)
    >
    > # now add density plot with histogram
    > check.overlap(x = x,
    + trt = trt01,
    + type = "both",
    + propensity.func = prop.func)
    >
    >
    > # simulated non-randomized treatment with multiple levels
    > xbetat_1 <- 0.15 + 0.5 * x[,9] - 0.25 * x[,12]
    > xbetat_2 <- 0.15 - 0.5 * x[,11] + 0.25 * x[,15]
    > trt.1.prob <- exp(xbetat_1) / (1 + exp(xbetat_1) + exp(xbetat_2))
    > trt.2.prob <- exp(xbetat_2) / (1 + exp(xbetat_1) + exp(xbetat_2))
    > trt.3.prob <- 1 - (trt.1.prob + trt.2.prob)
    > prob.mat <- cbind(trt.1.prob, trt.2.prob, trt.3.prob)
    > trt <- apply(prob.mat, 1, function(rr) rmultinom(1, 1, prob = rr))
    > trt <- apply(trt, 2, function(rr) which(rr == 1))
    >
    > # use multinomial logistic regression model with lasso penalty for propensity
    > propensity.multinom.lasso <- function(x, trt)
    + {
    + if (!is.factor(trt)) trt <- as.factor(trt)
    + gfit <- cv.glmnet(y = trt, x = x, family = "multinomial")
    +
    + # predict returns a matrix of probabilities:
    + # one column for each treatment level
    + propens <- drop(predict(gfit, newx = x, type = "response", s = "lambda.min",
    + nfolds = 5, alpha = 0))
    +
    + # return the probability corresponding to the
    + # treatment that was observed
    + probs <- propens[,match(levels(trt), colnames(propens))]
    +
    + probs
    + }
    >
    > check.overlap(x = x,
    + trt = trt,
    + type = "histogram",
    + propensity.func = propensity.multinom.lasso)
    Error in apply(nz, 1, median) : dim(X) must have a positive length
    Calls: check.overlap -> propensity.func -> cv.glmnet -> apply
    Execution halted
Flavor: r-patched-solaris-x86

Version: 0.1.3
Check: tests
Result: ERROR
     Running ‘testthat.R’ [243s/276s]
    Running the tests in ‘tests/testthat.R’ failed.
    Complete output:
     > Sys.setenv("R_TESTS" = "")
     > library(testthat)
     > library(personalized)
     Loading required package: glmnet
     Loading required package: Matrix
     Loading required package: foreach
     Loaded glmnet 2.0-13
    
     Loading required package: mgcv
     Loading required package: nlme
     This is mgcv 1.8-23. For overview type 'help("mgcv-package")'.
     Loading required package: gbm
     Loading required package: survival
     Loading required package: lattice
     Loading required package: splines
     Loading required package: parallel
     Loaded gbm 2.1.3
     Loading required package: ggplot2
     Loading required package: plotly
    
     Attaching package: 'plotly'
    
     The following object is masked from 'package:ggplot2':
    
     last_plot
    
     The following object is masked from 'package:stats':
    
     filter
    
     The following object is masked from 'package:graphics':
    
     layout
    
     >
     > test_check("personalized")
     ── 1. Error: test plot is returned for hist/density/both (@test-checkoverlap.R#1
     dim(X) must have a positive length
     1: check.overlap(x = x, trt = trt, type = "histogram", propensity.func = propensity.multinom.lasso) at testthat/test-checkoverlap.R:117
     2: propensity.func(x = x, trt = trt)
     3: cv.glmnet(y = trt, x = x, family = "multinomial") at testthat/test-checkoverlap.R:103
     4: apply(nz, 1, median)
     5: stop("dim(X) must have a positive length")
    
     ── 2. Error: test fit.subgroup for continuous outcomes and multiple trts and var
     dim(X) must have a positive length
     1: fit.subgroup(x = x, y = y, trt = trt, propensity.func = propensity.multinom.lasso,
     loss = "sq_loss_lasso", nfolds = 5) at testthat/test-fitsubgroup.R:1202
     2: propensity.func(x = x, trt = trt)
     3: cv.glmnet(y = trt, x = x, family = "multinomial") at testthat/test-fitsubgroup.R:1170
     4: apply(nz, 1, median)
     5: stop("dim(X) must have a positive length")
    
     ══ testthat results ═══════════════════════════════════════════════════════════
     OK: 580 SKIPPED: 0 FAILED: 2
     1. Error: test plot is returned for hist/density/both (@test-checkoverlap.R#117)
     2. Error: test fit.subgroup for continuous outcomes and multiple trts and various losses (@test-fitsubgroup.R#1202)
    
     Error: testthat unit tests failed
     Execution halted
Flavor: r-patched-solaris-x86