simulations/plot_dv_example.R

   1 source("RemembR/R/RemembeR.R")
   2 library(arrow)
   3 library(data.table)
   4 library(ggplot2)
   5 library(filelock)
   6 library(argparser)
   7
   8 parser <- arg_parser("Simulate data and fit corrected models.")
   9 parser <- add_argument(parser, "--infile", default="example_4.feather", help="name of the file to read.")
  10 parser <- add_argument(parser, "--remember-file", default="remembr.RDS", help="name of the remember file.")
  11 parser <- add_argument(parser, "--name", default="", help="The name to safe the data to in the remember file.")
  12 args <- parse_args(parser)
  13
  14 ## summarize.estimator <- function(df, suffix='naive', coefname='x'){
  15
  16 ##     part <- df[,c('N',
  17 ##                   'm',
  18 ##                   'Bxy',
  19 ##                   paste0('B',coefname,'y.est.',suffix),
  20 ##                   paste0('B',coefname,'y.ci.lower.',suffix),
  21 ##                   paste0('B',coefname,'y.ci.upper.',suffix),
  22 ##                   'y_explained_variance',
  23 ##                   'Bzy'
  24 ##                   ),
  25 ##                with=FALSE]
  26
  27 ##     true.in.ci <- as.integer((part$Bxy >= part[[paste0('B',coefname,'y.ci.lower.',suffix)]]) & (part$Bxy <= part[[paste0('B',coefname,'y.ci.upper.',suffix)]]))
  28 ##     zero.in.ci <- as.integer(0 >= part[[paste0('B',coefname,'y.ci.lower.',suffix)]]) & (0 <= part[[paste0('B',coefname,'y.ci.upper.',suffix)]])
  29 ##     bias <- part$Bxy - part[[paste0('B',coefname,'y.est.',suffix)]]
  30 ##     sign.correct <- as.integer(sign(part$Bxy) == sign(part[[paste0('B',coefname,'y.est.',suffix)]]))
  31
  32 ##     part <- part[,':='(true.in.ci = true.in.ci,
  33 ##                        zero.in.ci = zero.in.ci,
  34 ##                        bias=bias,
  35 ##                        sign.correct =sign.correct)]
  36
  37 ##     part.plot <- part[, .(p.true.in.ci = mean(true.in.ci),
  38 ##                           mean.bias = mean(bias),
  39 ##                           mean.est = mean(.SD[[paste0('B',coefname,'y.est.',suffix)]]),
  40 ##                           var.est = var(.SD[[paste0('B',coefname,'y.est.',suffix)]]),
  41 ##                           est.upper.95 = quantile(.SD[[paste0('B',coefname,'y.est.',suffix)]],0.95),
  42 ##                           est.lower.95 = quantile(.SD[[paste0('B',coefname,'y.est.',suffix)]],0.05),
  43 ##                           N.sims = .N,
  44 ##                           p.sign.correct = mean(as.integer(sign.correct & (! zero.in.ci))),
  45 ##                           variable=coefname,
  46 ##                           method=suffix
  47 ##                           ),
  48 ##                       by=c("N","m",'Bzy','y_explained_variance')
  49 ##                       ]
  50
  51 ##     return(part.plot)
  52 ## }
  53
  54 source("summarize_estimator.R")
  55
  56 build_plot_dataset <- function(df){
  57
  58     x.true <- summarize.estimator(df, 'true','x')
  59     z.true <- summarize.estimator(df, 'true','z')
  60
  61     x.naive <- summarize.estimator(df, 'naive','x')
  62     z.naive <- summarize.estimator(df, 'naive','z')
  63
  64     x.feasible <- summarize.estimator(df, 'feasible','x')
  65     z.feasible <- summarize.estimator(df, 'feasible','z')
  66
  67     x.amelia.full <- summarize.estimator(df, 'amelia.full','x')
  68     z.amelia.full <- summarize.estimator(df, 'amelia.full','z')
  69
  70     x.mle <- summarize.estimator(df, 'mle','x')
  71     z.mle <- summarize.estimator(df, 'mle','z')
  72
  73     x.zhang <- summarize.estimator(df, 'zhang','x')
  74     z.zhang <- summarize.estimator(df, 'zhang','z')
  75
  76     accuracy <- df[,mean(accuracy)]
  77
  78     plot.df <- rbindlist(list(x.true, z.true, x.naive,z.naive,x.amelia.full,z.amelia.full,x.mle, z.mle, x.zhang, z.zhang, x.feasible, z.feasible),use.names=T)
  79
  80     plot.df[,accuracy := accuracy]
  81
  82     plot.df <- plot.df[,":="(sd.est=sqrt(var.est)/N.sims)]
  83
  84     return(plot.df)
  85 }
  86
  87 change.remember.file(args$remember_file, clear=TRUE)
  88 sims.df <- read_feather(args$infile)
  89 sims.df[,Bzx:=NA]
  90 sims.df[,y_explained_variance:=NA]
  91 sims.df[,accuracy_imbalance_difference:=NA]
  92 plot.df <- build_plot_dataset(sims.df)
  93
  94 remember(plot.df,args$name)
  95
  96 set.remember.prefix(gsub("plot.df.","",args$name))
  97
  98 remember(median(sims.df$cor.xz),'med.cor.xz')
  99 remember(median(sims.df$accuracy),'med.accuracy')
 100 remember(median(sims.df$error.cor.x),'med.error.cor.x')
 101 remember(median(sims.df$error.cor.z),'med.error.cor.z')
 102 remember(median(sims.df$lik.ratio),'med.lik.ratio')
 103
 104
 105
 106
 107 ## df[gmm.ER_pval<0.05]
 108 ## plot.df.test <- plot.df[,':='(method=factor(method,levels=c("Naive","Multiple imputation", "Multiple imputation (Classifier features unobserved)","Regression Calibration","2SLS+gmm","Bespoke MLE", "Feasible"),ordered=T),
 109 ##                                    N=factor(N),
 110 ##                                    m=factor(m))]
 111
 112
 113 ## plot.df.test <- plot.df.test[(variable=='z') & (m != 1000) & (m!=500) & !is.na(p.true.in.ci) & (method!="Multiple imputation (Classifier features unobserved)")]
 114 ## p <- ggplot(plot.df.test, aes(y=mean.est, ymax=mean.est + var.est/2, ymin=mean.est-var.est/2, x=method))
 115 ## p <- p + geom_hline(aes(yintercept=-0.05),linetype=2)
 116
 117 ## p <- p + geom_pointrange() + facet_grid(m~N,as.table=F) + scale_x_discrete(labels=label_wrap_gen(4))
 118 ## print(p)
 119 ## ggplot(plot.df[variable=='x'], aes(y=mean.est, ymax=mean.est + var.est/2, ymin=mean.est-var.est/2, x=method)) + geom_pointrange() + facet_grid(-m~N) + scale_x_discrete(labels=label_wrap_gen(10))
 120
 121 ## ggplot(plot.df,aes(y=N,x=m,color=p.sign.correct)) + geom_point() + facet_grid(variable ~ method) + scale_color_viridis_c(option='D') + theme_minimal() + xlab("Number of gold standard labels") + ylab("Total sample size")
 122
 123 ## ggplot(plot.df,aes(y=N,x=m,color=abs(mean.bias))) + geom_point() + facet_grid(variable ~ method) + scale_color_viridis_c(option='D') + theme_minimal() + xlab("Number of gold standard labels") + ylab("Total sample size")