Last updated: 2017-01-15
Code version: 92038b69c4d891ab8162d54630d341956dc6e573
First, we load the necessary libraries.
library(ggplot2)
library(dplyr)
##
## Attaching package: 'dplyr'
## The following objects are masked from 'package:stats':
##
## filter, lag
## The following objects are masked from 'package:base':
##
## intersect, setdiff, setequal, union
Load and summarize the results of the data simulations.
load("../output/dsc-znull-files/res.znull.RData")
out <- ungroup(res %>% group_by(scenario,method) %>%
summarise(gt0 = mean(logLR > 0)))
xtabs(gt0 ~ method + scenario,out)
scenario
method znull.100 znull.1000
ash.hu 0.440 0.748
ash.hu.nw2 0.814 0.924
ash.hu.s 0.938 0.912
ash.hu.s.gmfine 0.932 0.908
ash.n 0.470 0.642
ash.n.nw2 0.570 0.736
ash.n.s 0.484 0.514
ash.n.s.gmfine 0.480 0.502
ash.u 0.214 0.428
ash.u.nw2 0.356 0.538
ash.u.s 0.504 0.556
ash.u.s.gmfine 0.502 0.554
Note that the ash.n.s conforms closely to the null expectation of 0.5 chi2_0 + 0.5 chi2_1 from Stram and Lee, Biometrics.
Conjecture: the uniform has the same asymptotic behaviour and the half uniform is like the sum of two of these?
From this a 95% procedure would need to check for logLR > 1.3528.
ggplot(res %>% filter(logLR > 0) %>% filter(grepl("ash.hu",method)),
aes(sample = logLR)) +
facet_grid(scenario ~ method) +
stat_qq(distribution = qchisq,dparams = list(df=1)) +
geom_abline() +
ggtitle("qqplot under null vs chisq-1; line slope=1")
ggplot(res %>% filter(logLR > 0) %>% filter(grepl("ash.u",method)),
aes(sample = logLR)) +
facet_grid(scenario~method) +
stat_qq(distribution = qchisq,dparams = list(df = 1)) +
geom_abline(intercept = 0,slope = 0.5) +
ggtitle("qqplot under null vs chisq-1; line slope=0.5")
ggplot(res %>% filter(logLR > 0) %>% filter(grepl("ash.n",method)),
aes(sample = logLR)) +
facet_grid(scenario ~ method) +
stat_qq(distribution = qchisq,dparams = list(df = 1)) +
geom_abline(intercept = 0,slope = 0.5) +
ggtitle("qqplot under null vs chisq-1; line slope=0.5")
sessionInfo()
R version 3.3.2 (2016-10-31)
Platform: x86_64-apple-darwin13.4.0 (64-bit)
Running under: macOS Sierra 10.12.2
locale:
[1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
attached base packages:
[1] stats graphics grDevices utils datasets methods base
other attached packages:
[1] knitr_1.15.1 dplyr_0.5.0 ggplot2_2.2.0
loaded via a namespace (and not attached):
[1] Rcpp_0.12.8 magrittr_1.5 munsell_0.4.3 xtable_1.8-2
[5] colorspace_1.2-6 R6_2.2.0 stringr_1.1.0 plyr_1.8.4
[9] tools_3.3.2 grid_3.3.2 gtable_0.2.0 DBI_0.5-1
[13] dscr_0.1.1 htmltools_0.3.5 yaml_2.1.14 lazyeval_0.2.0
[17] rprojroot_1.1 digest_0.6.10 assertthat_0.1 tibble_1.2
[21] shiny_0.14.2 reshape2_1.4.2 mime_0.5 evaluate_0.10
[25] rmarkdown_1.3 labeling_0.3 stringi_1.1.2 scales_0.4.1
[29] backports_1.0.4 httpuv_1.3.3