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
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
|
segreg <- function(formula, data, subset, weights, na.action, family=lm, control=seg.control(),
transf=NULL, contrasts=NULL, model=TRUE, x=FALSE, var.psi=TRUE, ...){
##### ====================================================================================
#DA FARE:
#1) i psi fissi OK prova
#2) eliminare argomento 'transf'
#3) se c'e' by consentire un n.psi diverso per categoria di by? ovvero npsi o psi devono essere liste..ok
#4) la matrice dei contrasti per imporre vincoli alle slope ok
#=================================
#Allora considerando seggrowth() qua ci dovrebbero essere problemi in nel predict..
#=================================
# `[.withAttributes` <- function(r, i) {
# subset <- NextMethod()
# attr(subset, "nomeBy") <- attr(r, "nomeBy")
# attr(subset, "nomeX") <- attr(r, "nomeX")
# attr(subset, "psi") <- attr(r, "psi")
# attr(subset, "npsi") <- attr(r, "npsi")
# attr(subset, "est") <- attr(r, "est")
# attr(subset, "R") <- attr(r, "R")
# attr(subset, "fix.psi") <- attr(r, "fix.psi")
# attr(subset, "f.x") <- attr(r, "f.x")
# attr(subset, "by") <- attr(r, "by")
# attr(subset, "levelsBy") <- attr(r, "levelsBy")
# subset
# }
build.all.psi<-function(psi, fixed.psi){
all.names.psi<-union(names(psi),names(fixed.psi))
all.psi<-vector("list", length=length(all.names.psi))
names(all.psi)<- all.names.psi
for(i in names(all.psi)) {
if(!is.null(psi[[i]])){
psi[[i]]<-sort(psi[[i]])
names(psi[[i]])<-paste("U",1:length(psi[[i]]),".",i,sep="")
}
if(!is.null(fixed.psi[[i]])){
fixed.psi[[i]]<-sort(fixed.psi[[i]])
names(fixed.psi[[i]])<- paste("U",1:length(fixed.psi[[i]]),".fixed.",i,sep="")
}
all.psi[[i]]<-sort(c(psi[[i]],fixed.psi[[i]]))
}
return(all.psi)
}
#=================================
fc<- min(max(abs(control$fc),.8),1)
min.step<-control$min.step
maxit.glm <- control$maxit.glm
alpha<-control$alpha
it.max <- old.it.max<- control$it.max
digits<-control$digits
toll <- control$toll
if(toll<0) stop("Negative tolerance ('tol' in seg.control()) is meaningless", call. = FALSE)
stop.if.error<-control$stop.if.error
fix.npsi<-fix.npsi<-control$fix.npsi
if(!is.null(stop.if.error)) {#if the old "stop.if.error" has been used..
warning(" Argument 'stop.if.error' is working, but will be removed in the next releases. Please use 'fix.npsi' for the future..")
} else {
stop.if.error<-fix.npsi
}
break.boot=control$break.boot
n.boot<-control$n.boot
size.boot<-control$size.boot
gap<-control$gap
random<-control$random
pow<-control$pow
conv.psi<-control$conv.psi
visual <- control$visual
visualBoot<-FALSE
if(visual && n.boot>0) {visual<-FALSE; visualBoot<-TRUE}
# if(n.boot>0){
# if(!is.null(control$seed)) {
# set.seed(control$seed)
# employed.Random.seed<-control$seed
# } else {
# employed.Random.seed<-eval(parse(text=paste(sample(0:9, size=6), collapse="")))
# set.seed(employed.Random.seed)
# }
# if(visual) {visual<-FALSE; visualBoot<-TRUE}# warning("`display' set to FALSE with bootstrap restart", call.=FALSE)}
# # if(!stop.if.error) stop("Bootstrap restart only with a fixed number of breakpoints")
# }
last <- control$last
K<-control$K
h<-control$h
if (deparse(substitute(family))=="lm" || (is.character(family) && family=="lm")){
fitter0<-"lm" #get("lm")
} else {
if (is.character(family)) family<-get(family, mode = "function", envir = parent.frame())
if (is.function(family)) family <- family()
if (is.null(family$family)) {
print(family)
stop("'family' not recognized")
}
fitter0<-"glm"
}
s1<-strsplit(as.character(formula)[3],"\\+")[[1]] #separa i termini "additivi"..
idC<-sapply(sapply(lapply(s1, function(x) grep("seg\\(",x)), function(x) (x>=1)), isTRUE)
stringa<-s1[idC] #solo i termini con seg
if(any(sapply(stringa, function(.x) grepl("\\* seg\\(", .x)))) stop("invalid usage of symbol '*' in conjunction with seg()")
if(any(sapply(stringa, function(.x) grepl("\\:seg\\(", .x)))) stop("invalid usage of symbol ':' in conjunction with seg()")
if(any(sapply(stringa, function(.x) grepl("\\):", .x)))) stop("invalid usage of symbol ':' in conjunction with seg()")
if(any(sapply(stringa, function(.x) grepl("\\) \\*", .x)))) stop("invalid usage of symbol '*' in conjunction with seg()")
call <- match.call()
if (missing(data)) data <- environment(formula)
tf <- terms(formula, specials = "seg")
id.ps<-attr(tf,"specials")$seg #posizione nel modelframe; vettore se ci sono piu' termini..include y ma non da interc
#browser()
mf <- match.call(expand.dots = FALSE)
m <- match(c("formula", "data", "subset", "weights", "na.action"), names(mf), 0L) #"offset",
mf <- mf[c(1, m)]
mf$drop.unused.levels <- TRUE
mf[[1L]] <- as.name("model.frame")
names(mf)[2]<-"formula" #serve se NON hai usato "formula"
#browser()
mf <- eval(mf, parent.frame())
#browser()
n<-nrow(mf)
mt <- attr(mf, "terms")
intercMt<-attr(mt,"intercept")
interc<-intercMt==1
Y <- model.response(mf, "any")
if (length(dim(Y)) == 1L) {
nm <- rownames(Y)
dim(Y) <- NULL
if (!is.null(nm)) names(Y) <- nm
}
if(!is.null(transf)) {
Y.orig <- Y
Y <- eval(parse(text=transf), list(y=Y))
transf.inv<-splinefun(Y, Y.orig, ties=min, method="monoH.FC")
}
if(is.null(alpha)) alpha<- max(.05, 1/length(Y))
if(length(alpha)==1) alpha<-c(alpha, 1-alpha)
#browser()
.xlivelli<-.getXlevels(mt, mf)
weights <- as.vector(model.weights(mf))
if(!is.null(weights) && !is.numeric(weights)) stop("'weights' must be a numeric vector")
if(!is.null(weights) && any(weights < 0)) stop("negative weights not allowed")
offs <- as.vector(model.offset(mf))
#browser() #funziona sia nella formula che come argomento?
testo.ps<-names(mf)[id.ps]
nomiCoefUNPEN<-names(mf)[-c(1,id.ps)]
# X <- if(!is.empty.model(mt)){
# model.matrix(mt, mf, contrasts)
# } else {stop("error in the design matrix")}#matrix(, NROW(Y), 0L)
# attrContr<-attr(X, "contrasts")
# n<-nrow(X)
#browser()
#===========================================================================
#se NON ci sono termini ps
#===========================================================================
if(length(testo.ps)<=0) stop("No seg() term. Please, use lm() or glm()")
#===========================================================================
#se ci sono termini ps.
#===========================================================================
check.estPsi<-function(.x){
#questo e' utile per verificare che estPsi sia o NA oppure di 0,1 (ma non tutti 0)
#l'ho fatta perche' se fosse consentito est=c(F,T) (e non FALSE,TRUE) poi T e F venivano presi come nomi di
# variabili in predict.segmented
if(length(.x)==1 && all(is.na(.x))){
ris<-TRUE
} else {
if(!is.numeric(.x)){
ris<-FALSE
} else {
if(length(setdiff(.x,0:1))==0){
ris<- if(all(.x==0)) FALSE else TRUE
} else {
ris<-FALSE
}
}
}
ris
}
drop.id<-lambda<-S<-B<-BB<-nomiCoefPEN<-nomiVariabPEN<-NULL
l<-lapply(testo.ps,function(xx)with(mf,eval(parse(text=xx),envir=data)))
nomiPS <-unlist(sapply(l,function(xx)attr(xx,"nomeX")))
psiList <-lapply(l,function(xx)attr(xx,"psi"))
#browser()
#npsiList <-unlist(sapply(l,function(xx) attr(xx,"npsi")))
npsiList <- lapply(l,function(xx) attr(xx,"npsi"))
estList <- lapply(l,function(xx) attr(xx,"est"))
if(length(setdiff(drop(unlist(sapply(estList, function(.x){if(any(is.na(.x))) 0 else unique(.x)}))), c(0,1)))>0) stop(" 'est' should include 0's and 1's only")
if(length(intersect(nomiCoefUNPEN, nomiPS))>0) stop("any segmented variable included as linear term too?")
#browser()
#if(!all(sapply(estList,check.estPsi))) stop(" 'est' is misspecified in one or more seg() term")
RList <- lapply(l,function(xx) attr(xx,"R"))
fixpsiList <- lapply(l,function(xx) attr(xx,"fix.psi"))
fxList <- lapply(l,function(xx) attr(xx,"f.x"))
byList <- lapply(l,function(xx) attr(xx,"by"))
nomiBy <- unlist(sapply(l,function(xx) attr(xx,"nomeBy")))
levelsBy <- lapply(l,function(xx) attr(xx,"levelsBy"))
#browser()
if(all(sapply(levelsBy, is.null)) && (length(npsiList)!=length(nomiPS))) stop(" 'npsi' is not correctly specified")
limZ<-rangeSmooth<-mVariabili<-NULL
#se ci sono termini ps()+ps(..,by) il nome delle variabili smooth vengono cambiati per aggiungere la variabile by
nomiPS.orig <- nomiPS
nomiPS.By <-paste(nomiPS, nomiBy, sep=":")
nomiPS <- unlist(lapply(nomiPS.By, function(.x) sub(":NULL", "", .x)))
####se la stessa variabile e' specificata come ps o termine lineare..
if(length(intersect(nomiPS,nomiCoefUNPEN))>=1) stop("The same variable specified in seg() and as linear term")
#ATTENZIONE.. se vuoi subito costruire i nomi ps(x), ps(x):z, ecc...usa i:
nomiPS.ps<- sapply(nomiPS.orig, function(.x)paste("seg(",list(.x),")",sep=""))
nomiPS.ps<-unlist(lapply(paste(nomiPS.ps, nomiBy, sep=":"), function(.x) sub(":NULL", "", .x)))
nomiPS.ps.list<-as.list(nomiPS.ps) #serve lista
for(j in id.ps) mVariabili[length(mVariabili)+1]<-mf[j]
B<- Bfix <- nomiPS.ps.int.list<-vector(length=length(mVariabili) , "list")
#BFixed<-BB<-Bderiv
#browser()
for(j in 1:length(mVariabili)) {
if(nomiBy[j]=="NULL"){ # se usuale termine seg()
nomiBy.j <- NULL
variabileSmooth<- c(mVariabili[[j]]) #c() converte le matrici in vettori, drop() no..!
#variabileSmooth<- attr(mVariabili[[j]], "f.x")(variabileSmooth)
variabileSmooth<- fxList[[j]](variabileSmooth)
#for(jj in c("nomeX", "psi", "npsi", "f.x", "nomeBy")) attr(variabileSmooth,jj)<-NULL
B[[j]]<- variabileSmooth
rangeSmooth[[j]] <- range(variabileSmooth)
limZ[[j]] <- quantile(variabileSmooth, names=FALSE, probs=c(alpha[1],alpha[2]))
nomiPS.ps.int.list[[j]]<- nomiPS[j]
} else { #se ci sono termini by
#browser()
if(is.null(levelsBy[[j]])){ #se e' vc con variabile continua
stop(" 'by' in seg(), if provided, should be a factor")
#B[[j]] <-variabileBy*B[[j]]
#nomiCoefPEN[[j]]<- sapply(1:ncol(B[[j]]), function(x) gsub(":", paste(".",x, ":", sep="") , nomiPS.ps[j]))
} else {#se e' VC con variabile categoriale
nomiBy.j <- nomiBy[j]
variabileSmooth<- mVariabili[[j]][,-ncol(mVariabili[[j]]),drop=TRUE]
variabileSmooth<- fxList[[j]](variabileSmooth)
variabileBy<- mVariabili[[j]][, ncol(mVariabili[[j]]),drop=TRUE]
M<-model.matrix(~0+factor(variabileBy))
B[[j]]<- lapply(1:ncol(M), function(.x) M[,.x]*variabileSmooth)
rangeSmooth[[j]] <- lapply(B[[j]], function(.x) range(.x[.x!=0]))
limZ[[j]] <- lapply(B[[j]], function(.x) quantile(.x[.x!=0], names=FALSE, probs=c(alpha[1],alpha[2])))
#browser()
cond1 <- is.list(psiList[[j]])
cond2 <- length(names(psiList[[j]]))==length(levelsBy[[j]])
cond3<- length(setdiff(names(psiList[[j]]),levelsBy[[j]]))==0
if(cond1&& cond2 && cond3) psiList[[j]]<- psiList[[j]][levelsBy[[j]]]
nomiPS.ps.list[[j]] <-paste(nomiPS.ps[j], levelsBy[[j]], sep="") #"seg(age):sex1" "seg(age):sex2"
nomiPS.ps.int.list[[j]]<-gsub("[)]", "", gsub("seg[(]", "", nomiPS.ps.list[[j]])) #age:sexM", "age:sexF"
#la linea sotto e' se hai diversi breakpoints nei gruppi..
}
}
} #end for(j in 1:length(mVariabili))
#browser()
#sapply(limZ[[1]], cbind)
repl<-pmax(sapply(B,length)*sapply(B,is.list),1)
for(i in 1:length(npsiList)){
if(length(npsiList[[i]])==1) {
npsiList[[i]] <- rep(npsiList[[i]], repl[i])
if(!is.list(estList[[i]]) && !is.null(levelsBy[[i]])) estList[[i]] <- rep(estList[i], repl[i])
}
if(length(nomiPS.ps.int.list[[i]])!=length(npsiList[[i]])) stop(paste(" 'npsi' (its length) is not correctly specified in the seg term:",i))
if(!is.null(names(npsiList[[i]]))){
if(length(setdiff(nomiPS.ps.int.list[[i]], names(npsiList[[i]])))!=0) stop(paste(" 'npsi' (its names) is not correctly specified in the seg term:",i))
}
}
npsiList <- unlist(npsiList)
#browser()
if(!any(sapply(psiList,is.list))) psiList <- rep(psiList, repl)
if(!any(sapply(estList,is.list))) estList <- rep(estList, repl)
if(!any(sapply(RList,is.list))) RList <- rep(RList, repl)
if(!any(sapply(fixpsiList,is.list))) fixpsiList<- rep(fixpsiList, repl)
nomiPS.orig <- rep(nomiPS.orig, repl)
Bfix <- rep(Bfix, repl)
while(any(sapply(B,is.list))){
id.vc<-which((sapply(B, is.list)))[1]
nc<-length(B[[id.vc]])
B<-append(B, B[[id.vc]], after = id.vc-1)
#for(i in 1:length(B[[id.vc+nc]])) BB<-append(BB, BB[id.vc], id.vc-1)
B[[id.vc+nc]]<-NULL
#BB[[id.vc+nc]]<-NULL
nomiCoefPEN <- append(nomiCoefPEN, nomiCoefPEN[[id.vc]], after = id.vc-1)
nomiCoefPEN[[id.vc+nc]]<-NULL
psiList <- append(psiList, psiList[[id.vc]], after = id.vc-1)
psiList[[id.vc+nc]]<-NULL
rangeSmooth <- append(rangeSmooth, rangeSmooth[[id.vc]], after = id.vc-1)
rangeSmooth[[id.vc+nc]]<-NULL
limZ <- append(limZ, limZ[[id.vc]], after = id.vc-1)
limZ[[id.vc+nc]]<-NULL
estList <- append(estList, estList[[id.vc]], after = id.vc-1)
estList[[id.vc+nc]]<-NULL
RList <- append(RList, RList[[id.vc]], after = id.vc-1)
RList[[id.vc+nc]]<-NULL
fixpsiList <- append(fixpsiList, fixpsiList[[id.vc]], after = id.vc-1)
fixpsiList[[id.vc+nc]]<-NULL
#se la lista contiene solo NULL, non funziona...
#penMatrixList <- append(penMatrixList, penMatrixList[[id.vc]], after = id.vc-1)
#penMatrixList[[id.vc+nc]]<-NULL
}
#if(!all(sapply(estList,check.estPsi))) stop(" 'est' is misspecified in one or more seg() term")
#browser()
nomiTerminiSEG<-nomiCoefPSI <-NULL
nomiPS.ps.unlist.seg <- unlist(nomiPS.ps.list)
nomiPS.ps.unlist <- sub("[)]", "", sub("seg[(]", "",nomiPS.ps.unlist.seg ))
names(psiList)<- nomiPS.ps.unlist
for(i in 1:length(B)) {
#nomiCoefPSI[[i]]<- paste(paste("psi",1:length(psiList[[i]]), sep=""), nomiPS.ps.unlist[i], sep=".") ##oppure sep=".psi"
nomiTerminiSEG[[i]]<-rep(nomiPS.ps.unlist[i], length(psiList[[i]]))
}
#nomiCoefU<-lapply(nomiCoefPSI, function(.x) sub("psi","U",.x ))
#nomiCoefZ<-lapply(nomiCoefPSI, function(.x) sub("psi","Z",.x ))
nomiSeg<- unique(unlist(nomiTerminiSEG))
#browser()
#FINALMENTE (speriamo..:-))
#nomiCoefZ, nomiCoefpsi, nomiCoefU sono liste con nomi che includono sia le possibili interazioni, sia il n. dei breakpoints
#Anche nomiTerminiSEG e' della stessa dimensione ma i nomi ignorano il n.dei breakpoints (questa serve per rangeZ)
#nomiSeg
#npsiList
#psiList
#estList
#RList
# if(is.null(names(estList))) {
# names(estList)<-nomiSeg
# } else {
# if(any(sapply(names(estList), function(.x).x==""))) stop(" 'estList' is only partially named.
# Or all or no name allowed.")
# }
#browser()
if(any(sapply(estList, is.list))) stop(" One or more 'est' components misspecified")
npsiList1<-id.contrR <- rep(NA, length(B))
psiListE<-psiListQ<-psiList
for(j in 1:length(B)){
#K<- npsiList[j]
K <- if(!is.na(npsiList[nomiSeg[j]])) npsiList[nomiSeg[j]] else npsiList[j]
npsiList1[j]<- K
if(any(is.na(psiList[[j]]))){
#if(control$quant) {
psiListQ[[j]]<- quantile(B[[j]], prob= seq(0,1,l=K+2)[-c(1,K+2)], names=FALSE)
#} else {
psiListE[[j]]<- (min(B[[j]])+ diff(range(B[[j]]))*(1:K)/(K+1))
#}
} else {
K<-npsiList1[j]<-length(psiList[[j]])
}
if(!is.null(fixpsiList[[j]])) {
Bfix[[j]]<- sapply(sort(fixpsiList[[j]]), function(.x) (B[[j]]-.x)*(B[[j]]>.x))
#colnames(Bfix[[j]])<- paste("U", 1:length(fixpsiList[[j]]),".fixed.",nomiPS.orig[j], sep="")
colnames(Bfix[[j]])<- paste("U", 1:length(fixpsiList[[j]]),".fixed.",nomiPS.ps.unlist[j], sep="")
}
#se per qualche termine ci sono le matrici dei vincoli sulle slope
#browser()
j.ok=match(nomiSeg[j], names(RList), nomatch=0)
j.ok <-if(j.ok>0) j.ok else j
if(!any(is.na(RList[[j.ok]]))){
RList[[j]] <- RList[[j.ok]]
id.contrR[j] <-TRUE
} else {
j.ok=match(nomiSeg[j], names(estList), nomatch=0)
j.ok <-if(j.ok>0) j.ok else j
if(!any(is.na(estList[[j.ok]]))){
if(length(estList[[j.ok]])!=(K+1)) stop(" 'est' is not compatible with 'npsi' ")
#browser()
RList[[j]]<-diag(K+1)[,estList[[j.ok]]==1,drop=FALSE]
id.contrR[j] <-TRUE
} else {
RList[[j]]<-diag(K+1)
id.contrR[j] <-FALSE
}
}
}
#browser()
if(control$quant) {
psiList<-psiListQ
initial <- unlist(psiListE)
PSI1<- matrix(initial, n, length(initial), byrow = TRUE)
} else {
psiList<-psiListE
initial <- unlist(psiListQ)
PSI1<- matrix(initial, n, length(initial), byrow = TRUE)
}
#Quindi PSI1 e' una matrice di valori inziali di psi.. Vanno usati da seg.lm.fit.boot nel caso in cui i primi PSI
#non portino ad adattare un modello
#NB: Poiche' ora psiList include i veri numeri dei psi, i codici vanno rilanciati
for(i in 1:length(B)) {
nomiCoefPSI[[i]]<- paste(paste("psi",1:length(psiList[[i]]), sep=""), nomiPS.ps.unlist[i], sep=".") ##oppure sep=".psi"
nomiTerminiSEG[[i]]<-rep(nomiPS.ps.unlist[i], length(psiList[[i]]))
}
nomiCoefU<-lapply(nomiCoefPSI, function(.x) sub("psi","U",.x ))
nomiCoefZ<-lapply(nomiCoefPSI, function(.x) sub("psi","Z",.x ))
npsii <- sapply(psiList,length)
id.psi.group <- rep(1:length(psiList), sapply(psiList,length))
Z<- lapply(1:length(B), function(.x) matrix(B[[.x]], nrow=n, ncol=npsiList1[[.x]]))
Z<- do.call(cbind,Z)
colnames(Z) <- unlist(nomiCoefZ)
#nomiPS, nomiPS.By, nomiPS.orig, nomiPS.ps, nomiPS.ps.list, nomiCateg, nomiInterCateg, nomiCoefPEN
#nomiPS: "x", "z" (vettore)
#nomiPS.orig: come "nomiPS"
#Se ci sono interazioni (by)
#nomiPS "x:g"
#nomiPS.orig: "x", "x", "x".. La stessa variabile ripetuta per il n.dei gruppi
#
#nomiPS.ps: "seg(x)", "seg(z)" (vettore) [con by: "seg(x):g"]
#nomiPS.ps.list "seg(x)", "seg(z)" (lista) [lista con by: "seg(x):g1" "seg(x):g2" "seg(x):g3" ..]
#nomiInterCateg: "x:g1" "x:g2" "x:g3" ..
#========================================================================================================
X <- if(!is.empty.model(mt)){
model.matrix(mt, mf, contrasts)
} else {stop("error in the design matrix")}#matrix(, NROW(Y), 0L)
attrContr<-attr(X, "contrasts")
#n<-nrow(X)
X<- X[, !startsWith(colnames(X),"seg("), drop=FALSE]
idZ <- unlist(tapply(id.psi.group, id.psi.group, function(.x) c(TRUE, rep(FALSE, length(.x)-1))))
Z.ok<-Z[, idZ, drop=FALSE]
colnames(Z.ok) <- nomiPS.ps.unlist
X<-cbind(X, Z.ok) #include anche i termini lineari delle variabili segmented
#colnames(Z)<- unlist(nomiCoefPEN)
initial <- unlist(psiList)
PSI <- matrix(initial, n, length(initial), byrow = TRUE)
#PSI1 di riserva definito sopra.
#browser()
#NB la matrie del disegno X include in nomi "seg(x)" e non va bene perche' poi da problemi con i
#nomi dei coef dell'oggetto.. Quindi bisogna sostituire questi nomi!!!
#non serve perche gia' i nomi sono ok..
#id.segX <-grep( "seg[(]" , colnames(X))
#colnames(X)[id.segX]<-gsub("[)]", "", gsub("seg[(]", "", colnames(X)[id.segX]))
if(any(!sapply(Bfix, is.null))){
X<-cbind(X, do.call(cbind, Bfix))
}
#browser()
#colnames(X)[unlist(id.psList)] <- nomiPS.orig
#X[,nomiPS.orig] <- Z[, unique(colnames(Z)), drop=FALSE]
#nomiCoefPEN include i nomi le interazioni con i livelli (nel caso vc) e anche del numero dei psi
#[1] "U1.x" "U2.x" "U1.z"
id.noOW <- if(is.null(weights) && is.null(offs)) TRUE else FALSE
if(is.null(weights)) weights<-rep(1,n)
orig.offs<-offs
if(is.null(offs)) offs<-rep(0,n)
#E' realmente necessario assegnare offs e weight con 0 e 1 anche se non servono???
#browser()
limZ <-do.call(cbind, lapply(limZ, function(.x){if(is.list(.x)) do.call(cbind, .x) else cbind(.x)} ))
#limZ <-matrix(sapply(1:length(npsii), function(.x) rep(limZ[,.x],npsii[.x])), nrow=2, byrow = FALSE)
limZ <- do.call(cbind, lapply(1:length(npsii), function(.x) matrix(limZ[,.x],nrow=2,ncol=npsii[.x])))
rangeZ <- do.call(cbind, lapply(1:length(npsii), function(.x) matrix(rangeSmooth[[.x]],nrow=2,ncol=npsii[.x])))
colnames(rangeZ)<-rep(names(npsii), npsii) #unlist(nomiTerminiSEG)
#browser()
#rangeZ<- matrix(sapply(1:length(npsii), function(.x) rep(rangeSmooth[[.x]],npsii[.x])), nrow=2, byrow = FALSE)
invXtX<-Xty<-NULL
#browser()
opz<-list(toll=toll,h=h,stop.if.error=stop.if.error,L0=NULL,visual=visual,it.max=it.max,nomiOK=unlist(nomiCoefU), usesegreg=TRUE,
fam=family, eta0=NULL, maxit.glm=maxit.glm, id.psi.group=id.psi.group, gap=gap, limZ=limZ, rangeZ=rangeZ,
conv.psi=conv.psi, alpha=alpha, fix.npsi=fix.npsi, min.step=min.step, tol.opt=control$tol.opt,
pow=pow, visualBoot=visualBoot, digits=digits, fc=fc, RList=RList, nomiSeg=nomiSeg,
seed=control$seed, min.n=control$min.n, PSI1=PSI1)
#browser()
if(any(id.contrR)){
if(fitter0=="lm"){
# for(.i in nomiSeg) { # #poni min(z)=0, cosi solve() in step.lm.fit non ha problemi.
# if(.i %in% colnames(X)) X[,.i]<- X[,.i] - min(X[,.i])
# }
if(n.boot <= 0) {
obj <- segConstr.lm.fit(Y, X, Z, PSI, weights, offs, opz)
} else {
obj <- segConstr.lm.fit.boot(Y, X, Z, PSI, weights, offs, opz,
n.boot = n.boot, size.boot = size.boot, random = random,
break.boot = break.boot)
# seed<- obj$seed
}
class0<- "lm"
if(obj$obj$df.residual==0) warning("no residual degrees of freedom (other warnings expected)", call.=FALSE)
} else {
if(n.boot<=0){
obj <-segConstr.glm.fit(Y, X, Z, PSI, weights, offs, opz)
} else {
obj <-segConstr.glm.fit.boot(Y, X, Z, PSI, weights, offs, opz,
n.boot=n.boot, size.boot=size.boot, random=random,
break.boot=break.boot)
# seed<- obj$seed
}
class0<-c("glm","lm")
}
} else {
#browser()
if(fitter0=="lm"){
if(n.boot <= 0) {
obj <- seg.lm.fit(Y, X, Z, PSI, weights, offs, opz)
} else {
#browser()
obj <- seg.lm.fit.boot(Y, X, Z, PSI, weights, offs, opz,
n.boot = n.boot, size.boot = size.boot, random = random,
break.boot = break.boot)
# seed<- obj$seed
}
class0<-"lm"
if(obj$obj$df.residual==0) warning("no residual degrees of freedom (other warnings expected)", call.=FALSE)
} else {
if(n.boot<=0){
obj <-seg.glm.fit(Y, X, Z, PSI, weights, offs, opz)
} else {
obj <-seg.glm.fit.boot(Y, X, Z, PSI, weights, offs, opz,
n.boot=n.boot, size.boot=size.boot, random=random,
break.boot=break.boot)
# seed<- obj$seed
}
class0<-c("glm","lm")
}
}
if(!is.list(obj)){
warning("Estimation failed. Too many breakpoints? Returning a (g)lm fit..", call. = FALSE)
if(fitter0=="lm"){
obj0 <- if(id.noOW) lm.fit(x = X, y = Y) else lm.wfit(x = X, y = Y, w = weights, offset = offs)
class(obj0)<-"lm"
} else {
obj0 <- try(suppressWarnings(glm.fit(X, y = Y, offset = offs,
weights = weights, family = opz$fam #control = glm.control(maxit = maxit.glm),
)), silent = TRUE)
class(obj0)<-c("glm", "lm")
}
return(obj0)
}
seed<- obj$seed
if(!is.list(obj)){
warning("No breakpoint estimated", call. = FALSE)
return(invisible(NULL))
}
#
id.psi.group<-obj$id.psi.group
npsi.groups <- tapply(id.psi.group, id.psi.group, length)
psi<-obj$psi
psi.values<-if(n.boot<=0) obj$psi.values else obj$boot.restart
U<-obj$U
V<-obj$V
rangeZ<-obj$rangeZ
#browser()
colnames(rangeZ) <- unlist(nomiTerminiSEG)
it <- obj$it
epsilon <- obj$epsilon
id.warn <- obj$id.warn
k <- length(psi)
objU <- obj$obj
#beta.c <- coef(objU)[paste("U", 1:ncol(U), sep = "")]
beta.c <- coef(objU)[obj$idU]
#browser()
if(any(id.contrR)) {
beta.c <- lapply(1:length(obj$constr$RList),
function(i) (obj$constr$invA.RList[[i]]%*%beta.c[unlist(obj$constr$nomiUList)==i])[-1])
beta.c <- unlist(beta.c)
nomiSeg <- rep(unique(unlist(nomiTerminiSEG)), sapply(obj$constr$nomiUList,function(.x) length(.x)))
replSeg <- unlist(sapply(obj$constr$nomiUList,function(.x) 1:length(.x)))
nomiU <- paste(paste("U", replSeg,sep=""), nomiSeg, sep=".")
nomiVxb <- sub("U","psi", obj$nomiOK)
X <- obj$X
} else {
nomiU <- obj$nomiOK #nomiOK ma puo' essere cambiato se sono eliminati dei psi nella procedura..
nomiVxb <- sub("U","psi", nomiU)
#In realta' nomiU e nomiVxb gia' ci sono (sarebbero nomiCoefU e nomiCoefPSI),
# pero' se durante la procedura sono stati cambiati perche' alcuni psi sono stati rimossi..bhu??
#nomiU <-nomiCoefU
#nomiVxb<- nomiCoefPSI
}
#browser()
Vxb <- V %*% diag(beta.c, ncol = length(beta.c))
colnames(U)<- nomiU #<- nomiOK
colnames(Vxb)<-nomiVxb #<- sub("U","psi", nomiU)
se.psi<-rep(NA,k)
if(fitter0=="lm"){
objV <- if(id.noOW) lm.fit(x = cbind(X, U, Vxb), y = Y) else lm.wfit(x = cbind(X, U, Vxb), y = Y, w = weights, offset = offs)
if(var.psi) {
s2 <- sum(weights*objU$residuals^2)/objV$df.residual
R <- chol2inv(objV$qr$qr)
se.psi <- sqrt(diag(R)*s2)[match(nomiVxb, names(coef(objV)),0)]
}
} else {
objV <- try(suppressWarnings(glm.fit(cbind(X, U, Vxb), y = Y, offset = offs,
weights = weights, family = opz$fam, #control = glm.control(maxit = maxit.glm),
etastart = objU$linear.predictors)), silent = TRUE)
objV$linear.predictors<-objU$linear.predictors
objV$deviance<-objU$deviance
objV$aic<-objU$aic + 2*ncol(PSI) #k
objV$weights<-objU$weights
#browser()
if (length(offs) && attr(mt, "intercept") > 0L) {
#se c'e' un offset devi calcolare la null.deviance (come fa glm())
obj0 <- try(suppressWarnings(glm.fit(X[, "(Intercept)", drop = FALSE], y = Y, offset = offs,
weights = weights, family = opz$fam, #control = glm.control(maxit = maxit.glm),
etastart = objU$linear.predictors, intercept=TRUE)), silent = TRUE)
# obj0 <- eval(call(if (is.function(method)) "method" else method,
# x = X[, "(Intercept)", drop = FALSE], y = Y, mustart = fit$fitted.values,
# weights = weights, offset = offset, family = family,
# control = control, intercept = TRUE))
if (!obj0$converged) warning("fitting to calculate the null deviance did not converge -- increase 'maxit'?")
objV$null.deviance <- obj0$deviance
}
if(var.psi) {
R <- chol2inv(objV$qr$qr)
s2 <- 1
if(!opz$fam$fam%in%c("poisson","binomial")) s2<- objU$deviance/objV$df.residual
se.psi <- sqrt(diag(R)*s2)[match(nomiVxb, names(coef(objV)),0)]
}
}
objV$fitted.values <- objU$fitted.values
objV$residuals <- objU$residuals
names.coef<-names(objV$coefficients)
objV$coefficients <- objU$coefficients
pLin<- ncol(X)
if(pLin>=1) {
names(objV$coefficients) <- c(names.coef[1:pLin], c(nomiU, nomiVxb))
} else {
names(objV$coefficients) <- c(nomiU, nomiVxb)
}
#browser()
ris.psi<-matrix(NA,k,3)
colnames(ris.psi) <- c("Initial", "Est.", "St.Err")
rownames(ris.psi) <- nomiVxb
ris.psi[,2]<-psi
ris.psi[,3]<- se.psi
if(stop.if.error) ris.psi[,1]<-initial
all.seg.form<-NULL
mf1<-mf[1]
for(i in 2:ncol(mf)) {
if(i %in% id.ps){
l<-attributes(mf[[i]])
if(!is.null(l$by)){
if(!l$nomeBy%in%names(mf)){
m<-data.frame(mf[[i]][,1],l$by)
colnames(m) <- c(l$nomeX, l$nomeBy)
} else {
m<-data.frame(mf[[i]][,1])
colnames(m) <- l$nomeX
}
all.seg.form[[length(all.seg.form)+1]]<-as.formula(
paste("~0+", l$nomeX, "*", l$nomeBy, "-", l$nomeX))
} else {
m <- data.frame(mf[[i]])
colnames(m) <- l$nomeX
all.seg.form[[length(all.seg.form)+1]]<- as.formula(paste("~", l$nomeX))
}
} else {
m <- mf[i]
}
mf1<-cbind(mf1, m)
}
#browser()
#costruisci la formulaLin.. Attenzione non tiene conto di eventuali vincoli sulle pendenze.
splitFo <- strsplit(as.character(formula),"[+]")
allX.lin<-paste(c(splitFo[[3]][-grep("seg[(]", splitFo[[3]])], unique(nomiPS.orig)), collapse="+")
formulaLin <- as.formula(paste(splitFo[[2]], splitFo[[1]], allX.lin))
#browser()
names(all.seg.form)<-nomiPS
#names(mf) <- nomi.mf
objV$terms <- mt
if(fitter0=="lm") objV$y<-Y #modificato il 6/5/24.. Non c'e' bisogno.. l'ogg restituit da glm.fit() ha gia' la y
if(x) objV$x <- X
objV$contrasts <- attrContr
objV$xlevels <- .xlivelli
objV$call<-call
#browser()
if (model) {
if(any(!sapply(levelsBy,is.null)) && any(id.contrR)) {
mf1<- cbind(mf1, Z.ok)
}
objV$model <- mf1
}
objV$na.action <- attr(mf, "na.action")
objV$psi <- ris.psi
objV$id.warn <- id.warn
objV$it <- it
objV$epsilon <- obj$epsilon
objV$rangeZ <- rangeZ
objV$constr <- obj$constr
objV$psi.history <- psi.values
#browser()
nomiPS.orig<-NULL
for(i in 1:length(nomiPS)){
nomiPS.orig[[i]]<-if(is.null(levelsBy[[i]])) nomiPS[i] else paste(nomiPS[i], unlist(levelsBy[[i]]), sep="")
}
nomiPS.orig <- unlist(nomiPS.orig)
#nomiPS.orig<-paste(nomiPS, unlist(levelsBy), sep="")
names(fixpsiList) <- nomiPS.orig
psi.list<-vector("list", length=length(unique(nomiPS.orig)))
#browser()
names(psi.list)<- nomiPS.orig
names(psi)<-rep(nomiPS.orig, npsi.groups)
for(i in names(psi.list)){
psi.list[[i]]<-psi[names(psi)==i]
}
objV$indexU<-build.all.psi(psi.list, fixpsiList)
#browser()
objV$nameUV <- list(U = drop(nomiU), V = rownames(ris.psi), Z = nomiPS.orig) #nomiPS.orig??
objV$nameUV$formulaSeg<- all.seg.form
objV$nameUV$formulaSegAllTerms<- paste("~", paste(sapply(all.seg.form, function(.x) strsplit(paste(.x), "~"))[2,],collapse="+"))
objV$formulaLin<- formulaLin
objV$id.psi.group<- id.psi.group
objV$psi[,"Initial"]<-NA
if(n.boot>0) objV$seed <- seed
#browser()
#il comando structure() l'ho messo perche' avevo bisogno che anche in mancanza di offset, l'oggetto finale restituisse
# un oggetto con la componente offset NULL. Cosa che non viene fatta con il semplice comando di sotto..
#objV<- if(id.O) c(objV, offset=offs) else c(objV, offset=NULL)
objV<- structure(c(objV, list(offset=orig.offs)))
class(objV)<-c("segmented", class0)
objV
}
|
