更新（基于r2evans＆＃39;评论）

Question

我在R中进行基于模拟的功率分析。我使用函数plyr::rdply和lme4::glmer运行R到RStudio（0.98.932）来生成数据并拟合模型，分别（参见下面R环境和包版本的可重现示例的结尾）。

该过程是随机生成给定参数化的数据集并使模型适合它。然而，时不时地，模型无法收敛。发生这种情况时会出现以下警告

[1]＆＃34;无法评估缩放的渐变＆＃34;
[2]＆＃34;模型未能收敛：使用1个负特征值退化Hessian＆＃34;

R进入浏览器模式，我必须手动干预（例如按c）返回模拟循环。这是一个真正的痛苦，因为我需要在几天内运行数千次迭代，但每次遇到这个特定的收敛错误时它都会停止，直到我按下一个键。

有没有办法避免R进入浏览器模式？我存储了每个模拟中出现的所有警告，因此我遇到的唯一问题是，当发生这种特定的收敛故障时，我必须手动干预。我尝试过使用purrr::quietly和purrr::safely函数，但没有成功（请参阅下面的代码示例）。

这是一个可以在我的计算机上运行的MWE（我使用set.seed进行再现性，所以我希望它可以独立于包版本等导致相同的结果）。在示例中，我使用相同的逻辑，但不同且更简单的参数化，正如我在实际模拟中使用的那样：

library(lme4)
library(plyr)
library(purrr)

# function to generate data that will lead to convergence failure
mini_simulator <- function() {
  nb_items <- 10  # observations per subject
  nb_subj <- 10  # subjects per group
  generate_data <- function() {
    A <- rbinom(nb_items * nb_subj, 1, .99)
    B <- rbinom(nb_items * nb_subj, 1, .8)
    simdata <- data.frame(
      Group = rep(c("A", "B"), each = nb_items * nb_subj),
      Subj = rep(1 : (nb_subj * 2), each = nb_items),
      Items = 1:nb_items,
      Response = c(A, B)
      )
  }
}

# Sanity check that the function is generating data appropriately.
# d should be a dataframe with 200 obs. of 4 variables
d <- mini_simulator()()
head(d, 3)
# Group Subj Items Response
# 1     A    1     1        1
# 2     A    1     2        1
# 3     A    1     3        1
rm(d)

## Functions to fit model

# basic function to fit model on simulated data
fit_model <- function(data_sim) {
  fm <- glmer(
    formula = Response ~ Group + (1|Subj) + (1|Items),
    data = data_sim, family = "binomial")
  out <- data.frame(summary(fm)$coef)
  out
}

# similar but using purrr::quietly (also tried purrr::safely with no success)
# see http://r4ds.had.co.nz/lists.html section "Dealing with failure"
fit_model_quietly <- function(data_sim) {
  purrr_out <- purrr::quietly(glmer)(
    formula = Response ~ Group + (1|Subj) + (1|Items),
    data = data_sim, family = "binomial")
  fm <- purrr_out$result
  out <- data.frame(summary(fm)$coef)
  # keeps track of convergence failures and other warnings
  out$Warnings <- paste(unlist(purrr_out$warnings), collapse = "; ")
  out
}

# this seed creates the problematic convergence failure on the first evaluation
# of rdply
set.seed(2)
# When I run the next line R goes into Browse mode and I need to enter "c"
# in order to continue
simulations <- plyr::rdply(.n = 3, fit_model(mini_simulator()()))
simulations

# problem persists using the quietly adverb from purrr
set.seed(2)
simulations <- plyr::rdply(.n = 3, fit_model_quietly(mini_simulator()()))
simulations

# sessionInfo()

# R version 3.1.2 (2014-10-31)
# Platform: i386-w64-mingw32/i386 (32-bit)
# 
# locale:
#   [1] LC_COLLATE=Swedish_Sweden.1252  LC_CTYPE=Swedish_Sweden.1252    LC_MONETARY=Swedish_Sweden.1252
# [4] LC_NUMERIC=C                    LC_TIME=Swedish_Sweden.1252    
# 
# attached base packages:
#   [1] stats     graphics  grDevices utils     datasets  methods   base     
# 
# other attached packages:
#   [1] purrr_0.2.1  plyr_1.8.1   lme4_1.1-8   Matrix_1.1-4
# 
# loaded via a namespace (and not attached):
#   [1] grid_3.1.2      lattice_0.20-29 magrittr_1.5    MASS_7.3-35     minqa_1.2.4     nlme_3.1-118    nloptr_1.0.4   
# [8] Rcpp_0.11.3     splines_3.1.2   tools_3.1.2

更新（基于r2evans＆＃39;评论）

在我的两台计算机options("error")上产生

（function（） { .rs.breakOnError（TRUE） }）（）

这似乎是某种RStudio的默认设置，实际上似乎让R在遇到stop()调用时进入浏览器模式（我看到这可以通过菜单工具栏Debug＆gt; On在图形界面中更改）错误＆gt; ...）。无论如何，当我设置options(error = NULL)时，问题就消失了。这是新的（简化的）示例，它可以很好地工作（在这个最小的例子中，也适用于实际的模拟）：

library(lme4)
library(plyr)
library(purrr)

options(error=NULL)

## Function to generate data
# Generates data that will lead to convergence failure
mini_simulator <- function() {
  nb_items <- 10  # observations per subject
  nb_subj <- 10  # subjects per group
  generate_data <- function() {
    A <- rbinom(nb_items * nb_subj, 1, .99)
    B <- rbinom(nb_items * nb_subj, 1, .8)
    simdata <- data.frame(
      Group = rep(c("A", "B"), each = nb_items * nb_subj),
      Subj = rep(1 : (nb_subj * 2), each = nb_items),
      Items = 1:nb_items,
      Response = c(A, B)
    )
  }
}

## Function to fit model
# Fits model on simulated data with purrr::quietly to capture warnings
# (http://r4ds.had.co.nz/lists.html section "Dealing with failure")
fit_model_quietly <- function(data_sim) {
  purrr_out <- purrr::quietly(glmer)(
    formula = Response ~ Group + (1|Subj) + (1|Items),
    data = data_sim, family = "binomial")
  fm <- purrr_out$result
  out <- data.frame(summary(fm)$coef)
  # keeps track of convergence failures and other warnings
  out$Warnings <- paste(unlist(purrr_out$warnings), collapse = "; ")
  out
}

# this seed creates the problematic convergence failure on the first evaluation
# of rdply
set.seed(2)
simulations <- plyr::rdply(.n = 3, fit_model_quietly(mini_simulator()()))
simulations

Answer 1

浏览器

这听起来像options的组合会触发警告和错误，以立即进入browser调试器。

func <- function(type = "none") {
  if (type == "warning") {
    warning("impending doom")
  } else if (type == "error") {
    stop("doom")    
  }
  type
}

func()
# [1] "none"
func("warning")
# Warning in func("warning") : impending doom
# [1] "warning"
func("error")
# Error in func("error") (from #5) : doom

两个相关选项，warn和error。有关详细信息，请参阅?options。

options("warn")
# $warn
# [1] 1

这会将警告转换为错误：

options(warn=2)
func("warning")
# Error in func("warning") (from #3) : (converted from warning) impending doom

现在使用error选项，我们可以对该错误采取措施：

options(warn=1, error=browser)
func("warning")
# Warning in func("warning") : impending doom
# [1] "warning"
func("error")
# Error in func("error") (from #5) : doom
# Browse[1]> 
c

因此将警告转换为错误并捕获错误：

options(warn=2, error=browser)
func("warning")
# Error in func("warning") (from #3) : (converted from warning) impending doom
# Browse[1]> 
c

我相信这就是发生在你身上的事。

`Purrr`

至于为什么purrr::quietly似乎做了一些事情，我可以确认它绕过或忽略了从警告到错误的预期升级：

quietfunc <- quietly(func)
str(quietfunc("warning"))
# List of 4
#  $ result  : chr "warning"
#  $ output  : chr ""
#  $ warnings: chr "impending doom"
#  $ messages: chr(0) 
options(warn=2, error=browser)
str(quietfunc("warning"))     # no browser!
# List of 4
#  $ result  : chr "warning"
#  $ output  : chr ""
#  $ warnings: chr "impending doom"
#  $ messages: chr(0) 
str(quietfunc("error"))       # yes browser
# Error in .f(...) (from #5) : doom
# Browse[1]> 
c

虽然safely没有抓住函数中的简单警告（也不打算这样做）：

options(warn=1)
str(safefunc("warning"))      # warning is not "captured" by purrr::safely
# Warning in .f(...) : impending doom
# List of 2
#  $ result: chr "warning"
#  $ error : NULL

也许是this is a bug？

为什么R会在收敛失败时自动进入浏览器模式？

更新（基于r2evans＆＃39;评论）

1 个答案:

浏览器

`Purrr`