R Under development (unstable) (2024-11-03 r87286 ucrt) -- "Unsuffered Consequences"
Copyright (C) 2024 The R Foundation for Statistical Computing
Platform: x86_64-w64-mingw32/x64

R is free software and comes with ABSOLUTELY NO WARRANTY.
You are welcome to redistribute it under certain conditions.
Type 'license()' or 'licence()' for distribution details.

R is a collaborative project with many contributors.
Type 'contributors()' for more information and
'citation()' on how to cite R or R packages in publications.

Type 'demo()' for some demos, 'help()' for on-line help, or
'help.start()' for an HTML browser interface to help.
Type 'q()' to quit R.

> library(sfsmisc)
> 
> ###--------------- "Iris Example for ever" ----------------------------
> data(iris)
> cl.true <- as.integer(iris[,"Species"])
> n <- length(cl.true)
> stopifnot(cl.true == rep(1:3, each = 50))
> m.iris <- data.matrix(iris[, 1:4])
> 
> .proctime00 <- proc.time()
> 
> ## Self Prediction:  Not too good (2+4 and 3+3 misclass.)
> table(diagDA(m.iris, cl.true, m.iris),             cl.true)
   cl.true
     1  2  3
  1 50  0  0
  2  0 48  4
  3  0  2 46
> table(diagDA(m.iris, cl.true, m.iris, pool=FALSE), cl.true)
   cl.true
     1  2  3
  1 50  0  0
  2  0 47  3
  3  0  3 47
> 
> ## Crossvalidation:  The same example as  knn() & knn1() from "class" :
> data(iris3)
> train <- rbind(iris3[1:25,,1], iris3[1:25,,2], iris3[1:25,,3])
> test <- rbind(iris3[26:50,,1], iris3[26:50,,2], iris3[26:50,,3])
> cl <- rep(1:3, each = 25)
> 
> pcl <- diagDA(train, cl, test)
> table(pcl, cl)## 0 + 1 + 2 misclassified
   cl
pcl  1  2  3
  1 25  0  0
  2  0 24  2
  3  0  1 23
> ## knn (    k=1) has 0 + 1 + 3
> ## knn ( *, k=3) has 0 + 2 + 3   ==> ``diagDA() is best ..''
> 
> stopifnot(pcl == diagDA(train,cl, test, pool = FALSE))
>                                         # i.e. quadratic identical here
> 
> ### Test 'NA' in predict dat.fr
> RNGversion("3.5.0")# -- so w/ sample() still stays unchanged:
Warning message:
In RNGkind("Mersenne-Twister", "Inversion", "Rounding") :
  non-uniform 'Rounding' sampler used
> set.seed(753)
> itr <- sample(n, 0.9 * n)
> lrn <- m.iris[ itr,]
> tst <- m.iris[-itr,]
> dd <- dDA(lrn, cl.true[itr])
> pd0 <- predict(dd, tst)
> 
> i.NA <- c(3:5,7,11)
> j.NA <- sample(1:ncol(tst), size=length(i.NA), replace=TRUE)
> tst[cbind(i.NA, j.NA)] <- NA
> pdd <- predict(dd, tst)
> pcl <- diagDA(lrn, cl.true[itr],  tst)
> stopifnot(length(pdd) == nrow(tst),
+           identical(pdd, pcl),
+           pdd[-i.NA] == pd0[-i.NA],
+           which(is.na(pdd)) == i.NA)
> 
> ## Now do some (randomized) CV :
> ## for each observation, count how often it's misclassified
> M <- 200
> set.seed(234)
> missCl <- integer(n)
> for(m in 1:M) {
+     itr <- sample(n, 0.9 * n)
+     lrn <- m.iris[ itr,]
+     tst <- m.iris[-itr,]
+     pcl <- diagDA(lrn, cl.true[itr],  tst)
+     stopifnot(pcl == predict(dDA(lrn, cl.true[itr]),  tst))
+     missCl <- missCl + as.integer(pcl != cl.true[ - itr])
+ }
> missCl ; mean(missCl) / M
  [1]  0  0  2  1  2  3  7 15 12 13  5 16 18 14  5  0  0  2  1  2  3  7 15 12 13
 [26]  5 16 18 14  5  0  0  2  1  2  3  7 15 12 13  5 16 18 14  5  0  0  2  1  2
 [51]  3  7 15 12 13  5 16 18 14  5  0  0  2  1  2  3  7 15 12 13  5 16 18 14  5
 [76]  0  0  2  1  2  3  7 15 12 13  5 16 18 14  5  0  0  2  1  2  3  7 15 12 13
[101]  5 16 18 14  5  0  0  2  1  2  3  7 15 12 13  5 16 18 14  5  0  0  2  1  2
[126]  3  7 15 12 13  5 16 18 14  5  0  0  2  1  2  3  7 15 12 13  5 16 18 14  5
[1] 0.03766667
> 
> ## The "same" with  'pool=FALSE' :
> missCl <- integer(n)
> for(m in 1:M) {
+     itr <- sample(n, 0.9 * n)
+     lrn <- m.iris[ itr,]
+     tst <- m.iris[-itr,]
+     pcl <- diagDA(lrn, cl.true[itr],  tst, pool=FALSE)
+     stopifnot(pcl == predict(dDA(lrn, cl.true[itr], pool=FALSE),  tst))
+     missCl <- missCl + as.integer(pcl != cl.true[ - itr])
+ }
> missCl ; mean(missCl) / M ## here somewhat worse than linear
  [1]  1  2  1  3  9 11 13  8 14 12 12 11 14 15  7  1  2  1  3  9 11 13  8 14 12
 [26] 12 11 14 15  7  1  2  1  3  9 11 13  8 14 12 12 11 14 15  7  1  2  1  3  9
 [51] 11 13  8 14 12 12 11 14 15  7  1  2  1  3  9 11 13  8 14 12 12 11 14 15  7
 [76]  1  2  1  3  9 11 13  8 14 12 12 11 14 15  7  1  2  1  3  9 11 13  8 14 12
[101] 12 11 14 15  7  1  2  1  3  9 11 13  8 14 12 12 11 14 15  7  1  2  1  3  9
[126] 11 13  8 14 12 12 11 14 15  7  1  2  1  3  9 11 13  8 14 12 12 11 14 15  7
[1] 0.04433333
> 
> cat('Time elapsed: ', proc.time() - .proctime00,'\n')
Time elapsed:  0.42 0 0.42 NA NA 
> 
> 
> proc.time()
   user  system elapsed 
   0.57    0.01    0.59