-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathSimulation.R
558 lines (512 loc) · 27.9 KB
/
Simulation.R
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
## ################################################
## Installing R-package 'ReconstPoFD' from GitHub
## ################################################
# install.packages("devtools")
# library("devtools")
# install_github("lidom/ReconstPoFD/ReconstPoFD")
## Load R-package 'ReconstPoFD'
library("ReconstPoFD")
## R-packages used for constructing the LaTeX-outputs:
library("tidyverse")
library("xtable")
## Location to store the simulation results:
your_path <- "your/location/here"
## #######################################
## Number of target functions
R <- 50
## Number of MC-Repetitions per target function
B <- 100
## #######################################
## Lower-value of the total domain
a <- 0
## Upper-value of the total domain
b <- 1
## #######################################
## Grid-size for evaluating the non-parametric
## estimates of the mean and the covariance
## For DGP1 (nRegGrid_vec[1]) we use a smaller
## value to speed up the computations
nRegGrid_vec <- c(21, 51)
## Number of binnings
## Smaller values speed up the computations
maxbins <- 100
## #######################################
## Sample sized
mm <- c(15, 30) # number of points per function
nn <- c(50, 100) # number of functions
## #######################################
## #######################################
## Simulate targets to be reconstructed
## in each simulation-step
## #######################################
Y_target_true_mat_LIST <- vector("list", c(4+4+2+2))
U_target_true_mat_LIST <- vector("list", c(4+4+2+2))
Y_target_mat_LIST <- vector("list", c(4+4+2+2))
U_target_mat_LIST <- vector("list", c(4+4+2+2))
Y_target_list_LIST <- vector("list", c(4+4+2+2))
U_target_list_LIST <- vector("list", c(4+4+2+2))
A_target_vec_LIST <- vector("list", c(4+4+2+2))
B_target_vec_LIST <- vector("list", c(4+4+2+2))
##
list_names <- c("DGP1_n50_m15","DGP1_n50_m30","DGP1_n100_m15","DGP1_n100_m30", # DGP1
"DGP2_n50_m15","DGP2_n50_m30","DGP2_n100_m15","DGP2_n100_m30", # DGP2
"DGP3_n50","DGP3_n100", # DGP3
"DGP4_n50","DGP4_n100") # DGP4
##
names(Y_target_true_mat_LIST) <- list_names
names(U_target_true_mat_LIST) <- list_names
names(Y_target_mat_LIST) <- list_names
names(U_target_mat_LIST) <- list_names
names(Y_target_list_LIST) <- list_names
names(A_target_vec_LIST) <- list_names
names(B_target_vec_LIST) <- list_names
for(DGP in c('DGP1','DGP2','DGP3','DGP4')){#
##
set.seed(1)
##
if(any(DGP==c("DGP1","DGP2"))){nRegGrid <- nRegGrid_vec[1]}
if(any(DGP==c("DGP3","DGP4"))){nRegGrid <- nRegGrid_vec[2]}
##
while(TRUE){
# Simulate data
SimDat <- ReconstPoFD::simuldata(n = R, m = mm[2], a = a, b = b, DGP=DGP, nRegGrid = nRegGrid)
if(!all(range(c(na.omit(SimDat$U_mat[,1])))==c(a,b))){
# takes only a non-degenerated partially observed function as a target function
break
}
}
##
if(any(DGP==c("DGP1","DGP2"))){
for(n in nn){
for(m in mm){ # n <- nn[1]; m <- mm[1]
if(m == mm[1]){slct <- sort(sample(1:mm[2], size = mm[1], replace = F))} # this selection guarantees that DGP1 and DGP2 have the same target functions, but only sampled at a different number of points m
if(m == mm[2]){slct <- 1:mm[2]}
Y_target_true_mat_LIST[[paste0(DGP,"_n",n,"_m",m)]] <- SimDat[['Y_true_mat']]
U_target_true_mat_LIST[[paste0(DGP,"_n",n,"_m",m)]] <- SimDat[['U_true_mat']]
Y_target_mat_LIST[[paste0(DGP,"_n",n,"_m",m)]] <- SimDat[['Y_mat']][slct,]
U_target_mat_LIST[[paste0(DGP,"_n",n,"_m",m)]] <- SimDat[['U_mat']][slct,]
for(r in 1:R){ # r <- 1
Y_target_list_LIST[[paste0(DGP,"_n",n,"_m",m)]][[r]] <- SimDat[['Y_list']][[r]][slct] # 'Y_mat' and 'Y_list' contain the same data!
U_target_list_LIST[[paste0(DGP,"_n",n,"_m",m)]][[r]] <- SimDat[['U_list']][[r]][slct] # 'U_mat' and 'U_list' contain the same data!
}
##
A_target_vec_LIST[[paste0(DGP,"_n",n,"_m",m)]] <- SimDat[['A_vec']]
B_target_vec_LIST[[paste0(DGP,"_n",n,"_m",m)]] <- SimDat[['B_vec']]
}
}
}
if(any(DGP==c("DGP3","DGP4"))){
for(n in nn){
Y_target_true_mat_LIST[[paste0(DGP,"_n",n)]] <- SimDat[['Y_true_mat']]
U_target_true_mat_LIST[[paste0(DGP,"_n",n)]] <- SimDat[['U_true_mat']]
Y_target_mat_LIST[[paste0(DGP,"_n",n)]] <- SimDat[['Y_mat']]
U_target_mat_LIST[[paste0(DGP,"_n",n)]] <- SimDat[['U_mat']]
for(r in 1:R){ # r <- 1
Y_target_list_LIST[[paste0(DGP,"_n",n)]][[r]] <- SimDat[['Y_list']][[r]]
U_target_list_LIST[[paste0(DGP,"_n",n)]][[r]] <- SimDat[['U_list']][[r]]
}
##
A_target_vec_LIST[[paste0(DGP,"_n",n)]] <- SimDat[['A_vec']]
B_target_vec_LIST[[paste0(DGP,"_n",n)]] <- SimDat[['B_vec']]
}
}
}
## ################################################################
## Start Simulation ##############################################
## ################################################################
(Start.Time <- Sys.time())
for(DGP in c('DGP1','DGP2','DGP3','DGP4')){
set.seed(2)
for(n in nn){
##
if(any(DGP==c("DGP1","DGP2"))){m_seq <- mm}
if(any(DGP==c("DGP3","DGP4"))){m_seq <- NA}
##
for(m in m_seq){
## DGP <- 'DGP1'; n <- nn[1]; m <- mm[1]; B <- 2; R <- 2
## #######################################################################
cat(DGP,"n=",n,"m=",m,"\n")
## #######################################################################
BiasSq_mat <- matrix(NA, nrow = R, ncol = 9)
Var_mat <- matrix(NA, nrow = R, ncol = 9)
if(any(DGP==c("DGP1","DGP2"))){
nRegGrid <- nRegGrid_vec[1]
slct_name <- paste0(DGP,"_n",n,"_m",m)
}
if(any(DGP==c("DGP3","DGP4"))){
nRegGrid <- nRegGrid_vec[2]
slct_name <- paste0(DGP,"_n",n)
}
##
for(r in 1:R){ # r <- 1
## ##################################################################
cat("r/R=",r,"/",R,"\n")
## ##################################################################
## Extract relevant targets
Y_target_true_mat <- Y_target_true_mat_LIST[[slct_name]][,r,drop=FALSE]
U_target_true_mat <- U_target_true_mat_LIST[[slct_name]][,r,drop=FALSE]
Y_target_mat <- Y_target_mat_LIST[[slct_name]][,r,drop=FALSE]
U_target_mat <- U_target_mat_LIST[[slct_name]][,r,drop=FALSE]
Y_target_list <- list("1"=Y_target_list_LIST[[slct_name]][[r]])
U_target_list <- list("1"=U_target_list_LIST[[slct_name]][[r]])
missings_of_target <- c(U_target_true_mat[,1] < A_target_vec_LIST[[slct_name]][r] |
U_target_true_mat[,1] > B_target_vec_LIST[[slct_name]][r])
## Declare variables for saving the results
Y_EYes_AYes <- matrix(NA, nrow = nRegGrid, ncol = B)
Y_EYes_ANo <- matrix(NA, nrow = nRegGrid, ncol = B)
Y_EYes_AYes_CES <- matrix(NA, nrow = nRegGrid, ncol = B)
Y_EYes_ANo_CES <- matrix(NA, nrow = nRegGrid, ncol = B)
Y_PACE <- matrix(NA, nrow = nRegGrid, ncol = B)
##
Y_ENo_CG_AYes <- matrix(NA, nrow = nRegGrid, ncol = B)
Y_ENo_CG_ANo <- matrix(NA, nrow = nRegGrid, ncol = B)
Y_Kraus <- matrix(NA, nrow = nRegGrid, ncol = B)
Y_PACE_ENo_CG <- matrix(NA, nrow = nRegGrid, ncol = B)
##
## #######################################################################
for(bb in 1:B){ # bb <- 1
##
while(TRUE){
SimDat <- ReconstPoFD::simuldata(n = n-1, m = m, a = a, b = b, DGP=DGP, nRegGrid = nRegGrid)
NonNA_fcts <- apply(SimDat[['Y_mat']],2,function(x)!any(is.na(x)))
n_NonNA_fcts <- length(NonNA_fcts[NonNA_fcts==TRUE])
##
if(n_NonNA_fcts >= floor(n*0.2)){
# guarantees that the samples contains at least 20% fully observed functions
break
}
}
##
Y_mat <- cbind(Y_target_mat, SimDat[['Y_mat']])
U_mat <- cbind(U_target_mat, SimDat[['U_mat']])
##
Y_list <- c(Y_target_list, SimDat[['Y_list']])
U_list <- c(U_target_list, SimDat[['U_list']])
##
if(any(DGP==c("DGP1", "DGP2"))){
##
## Reconstruction operator for noisy fragments with alignment at pre-smoothed fragment
result_EYes_AYes <- ReconstPoFD::reconstructKneipLiebl(Ly = Y_list,
Lu = U_list,
method = 'Error>0_AlignYES',
reconst_fcts = 1,
nRegGrid = nRegGrid,
maxbins = maxbins)
Y_EYes_AYes[,bb] <- matrix(unlist(result_EYes_AYes[['Y_reconst_list']]), nrow = nRegGrid, ncol = 1)
##
## Reconstruction operator for noisy fragments without alignment
result_EYes_ANo <- ReconstPoFD::reconstructKneipLiebl(Ly = Y_list,
Lu = U_list,
method = 'Error>=0_AlignNO',
reconst_fcts = 1,
nRegGrid = nRegGrid,
maxbins = maxbins)
Y_EYes_ANo[,bb] <- matrix(unlist(result_EYes_ANo[['Y_reconst_list']]), nrow = nRegGrid, ncol = 1)
##
## Reconstruction operator for noisy fragments with alignment at pre-smoothed fragment
result_EYes_AYes_CES <- ReconstPoFD::reconstructKneipLiebl(Ly = Y_list,
Lu = U_list,
method = 'Error>0_AlignYES_CEscores',
reconst_fcts = 1,
nRegGrid = nRegGrid,
maxbins = maxbins)
Y_EYes_AYes_CES[,bb] <- matrix(unlist(result_EYes_AYes_CES[['Y_reconst_list']]), nrow = nRegGrid, ncol = 1)
##
## Reconstruction operator for noisy fragments without alignment
result_EYes_ANo_CES <- ReconstPoFD::reconstructKneipLiebl(Ly = Y_list,
Lu = U_list,
method = 'Error>0_AlignNO_CEscores',
reconst_fcts = 1,
nRegGrid = nRegGrid,
maxbins = maxbins)
Y_EYes_ANo_CES[,bb] <- matrix(unlist(result_EYes_ANo_CES[['Y_reconst_list']]), nrow = nRegGrid, ncol = 1)
##
## PACE of Yao, Mueller, Wang (2005, JASA)
result_PACE <- ReconstPoFD::reconstructKneipLiebl(Ly = Y_list,
Lu = U_list,
method = 'PACE',
reconst_fcts = 1,
nRegGrid = nRegGrid,
maxbins = maxbins)
Y_PACE[,bb] <- matrix(unlist(result_PACE[['Y_reconst_list']]), nrow = nRegGrid, ncol = 1)
}
if(any(DGP==c("DGP3", "DGP4"))){
##
## Reconstruction operator for fully observed fragments with alignment at fully observed fragment
result_ENo_CG_AYes <- ReconstPoFD::reconstructKneipLiebl(Ly = Y_list,
Lu = U_list,
method = 'Error=0_AlignYES_CommonGrid',
reconst_fcts = 1,
nRegGrid = nRegGrid)
Y_ENo_CG_AYes[,bb] <- matrix(unlist(result_ENo_CG_AYes[['Y_reconst_list']]), nrow = nRegGrid, ncol = 1)
##
## Reconstruction operator for fully observed fragments without alignment
result_ENo_CG_ANo <- ReconstPoFD::reconstructKneipLiebl(Ly = Y_list,
Lu = U_list,
method = 'Error>=0_AlignNO',
reconst_fcts = 1,
nRegGrid = nRegGrid)
Y_ENo_CG_ANo[,bb] <- matrix(unlist(result_ENo_CG_ANo[['Y_reconst_list']]), nrow = nRegGrid, ncol = 1)
##
## Reconstruction Operator of Kraus (2015, JRSSB)
result_Kraus <- ReconstPoFD::reconstructKraus(X_mat = Y_mat, reconst_fcts = 1)
Y_Kraus[,bb] <- result_Kraus[['X_reconst_mat']]
##
## PACE for error==0
result_PACE_E0 <- ReconstPoFD::reconstructKneipLiebl(Ly = Y_list,
Lu = U_list,
method = 'Error=0_PACE',
reconst_fcts = 1,
nRegGrid = nRegGrid)
Y_PACE_ENo_CG[,bb] <- matrix(unlist(result_PACE_E0[['Y_reconst_list']]), nrow = nRegGrid, ncol = 1)
}
## ##################################################################
if(bb %% 50 == 0) cat("bb/B=",bb,"/",B,"\n")
## ##################################################################
} ## End of B-loop
##
slct_M <- missings_of_target
##
plotting <- FALSE
# plotting <- TRUE
if(plotting){
for(i in 1:min(B,10)){
if(any(DGP==c("DGP1", "DGP2"))){
par(mfrow=c(2,3))
##
plot(Y_EYes_AYes[,i], type="b", ylim=range(Y_EYes_AYes[,i],Y_target_true_mat[slct_M,1]),main="AYes")
lines(Y_target_true_mat[,1]); points(y=Y_EYes_AYes[slct_M,i], x=c(1:nRegGrid)[slct_M], col="red")
##
plot(Y_EYes_ANo[,i], type="b", ylim=range(Y_EYes_ANo[,i],Y_target_true_mat[slct_M,1]),main="ANo")
lines(Y_target_true_mat[,1]); points(y=Y_EYes_ANo[slct_M,i], x=c(1:nRegGrid)[slct_M], col="red")
##
plot(Y_EYes_AYes_CES[,i], type="b", ylim=range(Y_EYes_AYes_CES[,i],Y_target_true_mat[slct_M,1]),main="AYesCE")
lines(Y_target_true_mat[,1]); points(y=Y_EYes_AYes_CES[slct_M,i], x=c(1:nRegGrid)[slct_M], col="red")
##
plot(Y_EYes_ANo_CES[,i], type="b", ylim=range(Y_EYes_ANo_CES[,i],Y_target_true_mat[slct_M,1]),main="ANoCE")
lines(Y_target_true_mat[,1]); points(y=Y_EYes_ANo_CES[slct_M,i], x=c(1:nRegGrid)[slct_M], col="red")
##
plot(Y_PACE[,i], type="b", ylim=range(Y_PACE[,i],Y_target_true_mat[slct_M,1]),main="PACE")
lines(Y_target_true_mat[,1]); points(y=Y_PACE[slct_M,i], x=c(1:nRegGrid)[slct_M], col="red")
}
##
if(any(DGP==c('DGP3','DGP4'))){
par(mfrow=c(2,2))
plot(Y_ENo_CG_AYes[,i], type="b", ylim=range(Y_ENo_CG_AYes[,i],Y_target_true_mat[slct_M,1]),main="AYes")
lines(Y_target_true_mat[,1]); points(y=Y_ENo_CG_AYes[slct_M,i], x=c(1:nRegGrid)[slct_M], col="red")
##
plot(Y_ENo_CG_ANo[,i], type="b", ylim=range(Y_ENo_CG_ANo[,i],Y_target_true_mat[slct_M,1]),main="ANo")
lines(Y_target_true_mat[,1]); points(y=Y_ENo_CG_ANo[slct_M,i], x=c(1:nRegGrid)[slct_M], col="red")
##
plot(Y_Kraus[,i], type="b", ylim=range(Y_Kraus[,i],Y_target_true_mat[slct_M,1]),main="KRAUS")
lines(Y_target_true_mat[,1]); points(y=Y_Kraus[slct_M,i], x=c(1:nRegGrid)[slct_M], col="red")
##
plot(Y_PACE_ENo_CG[,i], type="b", ylim=range(Y_PACE_ENo_CG[,i],Y_target_true_mat[slct_M,1]),main="PACE")
lines(Y_target_true_mat[,1]); points(y=Y_PACE_ENo_CG[slct_M,i], x=c(1:nRegGrid)[slct_M], col="red")
}
par(mfrow=c(1,1))
Sys.sleep(1)
}
}
##
cut <- 0.005 # (if cut=0, no winsorization)
##
Y_EYes_AYes_Mean <- apply(Y_EYes_AYes[ slct_M, ], 1, function(x){mean(ReconstPoFD:::winsorize_x(x,cut=cut))})
Y_EYes_ANo_Mean <- apply(Y_EYes_ANo[ slct_M, ], 1, function(x){mean(ReconstPoFD:::winsorize_x(x,cut=cut))})
Y_EYes_AYes_CES_Mean <- apply(Y_EYes_AYes_CES[slct_M, ], 1, function(x){mean(ReconstPoFD:::winsorize_x(x,cut=cut))})
Y_EYes_ANo_CES_Mean <- apply(Y_EYes_ANo_CES[ slct_M, ], 1, function(x){mean(ReconstPoFD:::winsorize_x(x,cut=cut))})
Y_PACE_Mean <- apply(Y_PACE[ slct_M, ], 1, function(x){mean(ReconstPoFD:::winsorize_x(x,cut=cut))})
##
Y_ENo_CG_AYes_Mean <- apply(Y_ENo_CG_AYes[ slct_M, ], 1, function(x){mean(ReconstPoFD:::winsorize_x(x,cut=cut))})
Y_ENo_CG_ANo_Mean <- apply(Y_ENo_CG_ANo[ slct_M, ], 1, function(x){mean(ReconstPoFD:::winsorize_x(x,cut=cut))})
Y_Kraus_Mean <- apply(Y_Kraus[ slct_M, ], 1, function(x){mean(ReconstPoFD:::winsorize_x(x,cut=cut))})
Y_PACE_ENo_Mean <- apply(Y_PACE_ENo_CG[ slct_M, ], 1, function(x){mean(ReconstPoFD:::winsorize_x(x,cut=cut))})
##
## ## \int_{Missing} (E(X(t))-X(t))^2 dt ('Integrated squared bias')
Y_EYes_AYes_Int_BiasSq <- sum( (Y_EYes_AYes_Mean - Y_target_true_mat[slct_M, 1])^2 ) * (b-a)/nRegGrid
Y_EYes_ANo_Int_BiasSq <- sum( (Y_EYes_ANo_Mean - Y_target_true_mat[slct_M, 1])^2 ) * (b-a)/nRegGrid
Y_EYes_AYes_CES_Int_BiasSq <- sum( (Y_EYes_AYes_CES_Mean - Y_target_true_mat[slct_M, 1])^2 ) * (b-a)/nRegGrid
Y_EYes_ANo_CES_Int_BiasSq <- sum( (Y_EYes_ANo_CES_Mean - Y_target_true_mat[slct_M, 1])^2 ) * (b-a)/nRegGrid
Y_PACE_Int_BiasSq <- sum( (Y_PACE_Mean - Y_target_true_mat[slct_M, 1])^2 ) * (b-a)/nRegGrid
##
Y_ENo_CG_AYes_Int_BiasSq <- sum( (Y_ENo_CG_AYes_Mean - Y_target_true_mat[slct_M, 1])^2 ) * (b-a)/nRegGrid
Y_ENo_CG_ANo_Int_BiasSq <- sum( (Y_ENo_CG_ANo_Mean - Y_target_true_mat[slct_M, 1])^2 ) * (b-a)/nRegGrid
Y_Kraus_Int_BiasSq <- sum( (Y_Kraus_Mean - Y_target_true_mat[slct_M, 1])^2 ) * (b-a)/nRegGrid
Y_PACE_ENo_Int_BiasSq <- sum( (Y_PACE_ENo_Mean - Y_target_true_mat[slct_M, 1])^2 ) * (b-a)/nRegGrid
## \int_{Missing} Var(X(t)) dt ('Integrated variance')
Y_EYes_AYes_Int_Var <- sum(apply(Y_EYes_AYes[ slct_M, ], 1, function(x){var(ReconstPoFD:::winsorize_x(x,cut=cut))})) * (b-a)/nRegGrid
Y_EYes_ANo_Int_Var <- sum(apply(Y_EYes_ANo[ slct_M, ], 1, function(x){var(ReconstPoFD:::winsorize_x(x,cut=cut))})) * (b-a)/nRegGrid
Y_EYes_AYes_CES_Int_Var <- sum(apply(Y_EYes_AYes_CES[slct_M, ], 1, function(x){var(ReconstPoFD:::winsorize_x(x,cut=cut))})) * (b-a)/nRegGrid
Y_EYes_ANo_CES_Int_Var <- sum(apply(Y_EYes_ANo_CES[ slct_M, ], 1, function(x){var(ReconstPoFD:::winsorize_x(x,cut=cut))})) * (b-a)/nRegGrid
Y_PACE_Int_Var <- sum(apply(Y_PACE[ slct_M, ], 1, function(x){var(ReconstPoFD:::winsorize_x(x,cut=cut))})) * (b-a)/nRegGrid
##
Y_ENo_CG_AYes_Int_Var <- sum(apply(Y_ENo_CG_AYes[ slct_M, ], 1, function(x){var(ReconstPoFD:::winsorize_x(x,cut=cut))})) * (b-a)/nRegGrid
Y_ENo_CG_ANo_Int_Var <- sum(apply(Y_ENo_CG_ANo[ slct_M, ], 1, function(x){var(ReconstPoFD:::winsorize_x(x,cut=cut))})) * (b-a)/nRegGrid
Y_Kraus_Int_Var <- sum(apply(Y_Kraus[ slct_M, ], 1, function(x){var(ReconstPoFD:::winsorize_x(x,cut=cut))})) * (b-a)/nRegGrid
Y_PACE_ENo_Int_Var <- sum(apply(Y_PACE_ENo_CG[ slct_M, ], 1, function(x){var(ReconstPoFD:::winsorize_x(x,cut=cut))})) * (b-a)/nRegGrid
##
BiasSq_mat[r,] <- c(Y_EYes_AYes_Int_BiasSq, Y_EYes_ANo_Int_BiasSq,
Y_EYes_AYes_CES_Int_BiasSq, Y_EYes_ANo_CES_Int_BiasSq,
Y_PACE_Int_BiasSq,
##
Y_ENo_CG_AYes_Int_BiasSq, Y_ENo_CG_ANo_Int_BiasSq, Y_Kraus_Int_BiasSq, Y_PACE_ENo_Int_BiasSq)
###
Var_mat[r,] <- c(Y_EYes_AYes_Int_Var, Y_EYes_ANo_Int_Var,
Y_EYes_AYes_CES_Int_Var, Y_EYes_ANo_CES_Int_Var,
Y_PACE_Int_Var,
##
Y_ENo_CG_AYes_Int_Var, Y_ENo_CG_ANo_Int_Var, Y_Kraus_Int_Var, Y_PACE_ENo_Int_Var)
}# end of r-loop
##
## Tacking the mean over all target functions
BiasSq_vec <- colMeans(BiasSq_mat)
Var_vec <- colMeans(Var_mat)
##
name_vec <- c("EYes_AYes", "EYes_ANo",
"EYes_AYes_CES", "EYes_ANo_CES",
"PACE",
##
"ENo_CG_AYes", "ENo_CG_ANo", "Kraus", "PACE_E0")
names(BiasSq_vec) <- name_vec
names(Var_vec) <- name_vec
##
## Save results:
if(any(DGP==c("DGP1", "DGP2"))){
save(BiasSq_vec, Var_vec, file = paste0(your_path,DGP,"_n",n,"_m",m,"_simResults.RData"))
}
if(any(DGP==c('DGP3','DGP4'))){
save(BiasSq_vec, Var_vec, file = paste0(your_path,DGP,"_n",n,"_simResults.RData"))
}
}
}
}
##------------------------------------
End.Time <- Sys.time()
## Run-time:
round(End.Time - Start.Time, 2)
##------------------------------------
## ######################################################
## Build LaTeX tables for the simulation results
## ######################################################
## --------------------------------
## DGP1
## --------------------------------
DGP1_df_0 <- NULL
for(n in c(50, 100)){
for(m in c(15, 30)){
## n <- 50; m <- 15
load(file = paste0(your_path,"DGP1_n",n,"_m",m,"_simResults.RData"))
##
col_slct <- !is.na(BiasSq_vec)
BiasSq_vec <- BiasSq_vec[col_slct]
Var_vec <- Var_vec[col_slct]
MSE_vec <- BiasSq_vec + Var_vec
##
DGP1_df_0 <- rbind.data.frame(DGP1_df_0,
tibble("DGP" = as.factor("DGP1"),
"Estim" = as.factor(c("EYes_AYes", "EYes_ANo",
"EYes_AYes_CES", "EYes_ANo_CES",
"PACE")),
"n" = as.integer(n),
"m" = as.integer(m),
"MSE" = MSE_vec,
"Bias2" = BiasSq_vec,
"Var" = Var_vec)
)
}
}
DGP1_df <- DGP1_df_0 %>%
group_by(n, m) %>%
arrange(n, m, MSE) %>%
mutate("MSEr" = MSE / min(MSE)) %>%
ungroup() %>%
mutate(Estim = fct_recode(Estim,
AYes = "EYes_AYes",
ANo = "EYes_ANo",
AYesCE = "EYes_AYes_CES",
ANoCE = "EYes_ANo_CES",
PACE = "PACE")) %>%
select(DGP, n, m, Estim, MSEr, MSE, Bias2, Var)
print(DGP1_df, n = 20)
DGP1_df_LaTeX <- xtable(DGP1_df, digits=c(0,0,0,0,0,2,3,3,3))
print(DGP1_df_LaTeX, include.rownames = FALSE)
## --------------------------------
## DGP2
## --------------------------------
DGP2_df_0 <- NULL
for(n in c(50, 100)){
for(m in c(15, 30)){
## n <- 50; m <- 15
load(file = paste0(your_path,"DGP2_n",n,"_m",m,"_simResults.RData"))
##
col_slct <- !is.na(BiasSq_vec)
BiasSq_vec <- BiasSq_vec[col_slct]
Var_vec <- Var_vec[col_slct]
MSE_vec <- BiasSq_vec + Var_vec
##
DGP2_df_0 <- rbind.data.frame(DGP2_df_0,
tibble("DGP" = as.factor("DGP2"),
"Estim" = as.factor(c("EYes_AYes", "EYes_ANo",
"EYes_AYes_CES", "EYes_ANo_CES",
"PACE")),
"n" = as.integer(n),
"m" = as.integer(m),
"MSE" = MSE_vec,
"Bias2" = BiasSq_vec,
"Var" = Var_vec)
)
}
}
DGP2_df <- DGP2_df_0 %>%
group_by(n, m) %>%
arrange(n, m, MSE) %>%
mutate("MSEr" = MSE / min(MSE)) %>%
ungroup() %>%
mutate(Estim = fct_recode(Estim,
AYes = "EYes_AYes",
ANo = "EYes_ANo",
AYesCE = "EYes_AYes_CES",
ANoCE = "EYes_ANo_CES",
PACE = "PACE")) %>%
select(DGP, n, m, Estim, MSEr, MSE, Bias2, Var)
print(DGP2_df, n = 20)
DGP2_df_LaTeX <- xtable(DGP2_df, digits=c(0,0,0,0,0,2,3,3,3))
print(DGP2_df_LaTeX, include.rownames = FALSE)
## --------------------------------
## DGP3 and DGP4
## --------------------------------
DGP3n4_df_0 <- NULL
for(DGP in c('DGP3', 'DGP4')){
for(n in c(50, 100)){
##
load(file = paste0(your_path,DGP,"_n",n,"_simResults.RData"))
##
col_slct <- !is.na(BiasSq_vec)
BiasSq_vec <- BiasSq_vec[col_slct]
Var_vec <- Var_vec[col_slct]
MSE_vec <- BiasSq_vec + Var_vec
##
DGP3n4_df_0 <- rbind.data.frame(DGP3n4_df_0,
tibble("DGP" = as.factor(DGP),
"Estim" = as.factor(c("ENo_CG_AYes", "ENo_CG_ANo", "Kraus", "PACE_E0")),
"n" = as.integer(n),
"MSE" = MSE_vec,
"Bias2" = BiasSq_vec,
"Var" = Var_vec)
)
}
}
DGP3n4_df <- DGP3n4_df_0 %>%
mutate(Estim = fct_recode(Estim,
AYes = "ENo_CG_AYes",
ANo = "ENo_CG_ANo",
KRAUS = "Kraus",
PACE = "PACE_E0")) %>%
group_by(DGP, n) %>%
arrange(DGP, n, MSE) %>%
mutate("MSEr" = MSE / min(MSE)) %>%
ungroup() %>%
select(DGP, n, Estim, MSEr, MSE, Bias2, Var)
DGP3n4_df
DGP2n3_df_LaTeX <- xtable(DGP2n3_df, digits=c(0,0,0,0,2,3,3,3))
print(DGP2n3_df_LaTeX, include.rownames = FALSE)