Tests on artificial data with 80% noise

library(adabag)
library(naivebayes)

Loading required package: rpart
Loading required package: caret
Loading required package: lattice
Loading required package: ggplot2
Loading required package: foreach
Loading required package: doParallel
Loading required package: iterators
Loading required package: parallel

# files available in: /net/aistaff/kleiweg/spraak/fa
train = read.table("data080.train", header=TRUE, sep="\t", quote="", row.names=1)
test  = read.table("data080.test",  header=TRUE, sep="\t", quote="", row.names=1)

train[1:10,]

	C.Class	C.W1	C.W2	C.W3	C.W4	C.W5	C.W6	C.W7	C.W8	C.W9	⋯	C.W11	C.W12	C.W13	C.W14	C.W15	C.W16	C.W17	C.W18	C.W19	C.W20
135	C	1.C1	2.B1	3.B1	4.C4	5.B5	6.A1	7.C1	8.B1	9.C7	⋯	11.A1	12.C1	13.C1	14.C2	15.C1	16.C1	17.C6	18.B1	19.C7	20.B1
488	C	1.B1	2.C1	3.B1	4.C4	5.C2	6.B1	7.C2	8.A1	9.B4	⋯	11.A1	12.C2	13.B1	14.A3	15.A1	16.C1	17.B1	18.A1	19.C7	20.A3
641	B	1.C1	2.B1	3.B1	4.B3	5.B2	6.A1	7.C1	8.A1	9.C6	⋯	11.B4	12.A1	13.B1	14.B1	15.A1	16.A1	17.A1	18.B6	19.B3	20.B1
885	B	1.B1	2.C1	3.A1	4.B1	5.B5	6.B1	7.A1	8.A1	9.B1	⋯	11.A1	12.A1	13.C4	14.C5	15.B3	16.C2	17.A1	18.B1	19.B6	20.B2
148	B	1.B1	2.B1	3.B1	4.B1	5.B4	6.B1	7.B1	8.C1	9.A2	⋯	11.B3	12.B1	13.A3	14.B1	15.C1	16.C1	17.B5	18.B5	19.A4	20.B1
576	B	1.B1	2.C1	3.C2	4.B3	5.B5	6.C2	7.B1	8.A1	9.B3	⋯	11.C2	12.A1	13.B2	14.B3	15.B5	16.A1	17.A1	18.C2	19.B4	20.A1
497	B	1.B1	2.B1	3.B1	4.B3	5.A1	6.C4	7.B1	8.B1	9.B4	⋯	11.C1	12.A1	13.A3	14.C5	15.C1	16.A3	17.A1	18.B5	19.A3	20.B2
97	B	1.C1	2.B2	3.C2	4.C2	5.A2	6.A1	7.C2	8.A1	9.B1	⋯	11.B3	12.A1	13.C4	14.A3	15.C1	16.A2	17.B1	18.A1	19.B3	20.A3
33	B	1.A5	2.A4	3.B1	4.A2	5.B5	6.B1	7.B1	8.C1	9.B3	⋯	11.B3	12.C2	13.B1	14.B1	15.C1	16.C1	17.C5	18.C2	19.B6	20.B1
183	A	1.A1	2.C1	3.C2	4.A2	5.A2	6.B1	7.A2	8.A1	9.A2	⋯	11.A1	12.A1	13.C5	14.A1	15.A3	16.C1	17.A1	18.B5	19.B4	20.B2

Bagging (AdaBag)

bag <- bagging(C.Class ~ ., data=train)
train.bagging <- predict(bag, newdata=train)
 test.bagging <- predict(bag, newdata=test)
100 * (1 - train.bagging$error)
100 * (1 -  test.bagging$error)

91.5555555555556

63

Boosting (AdaBoost)

boost <- boosting(C.Class ~ ., data=train)
train.boosting <- predict(boost, newdata=train)
 test.boosting <- predict(boost, newdata=test)
100 * (1 - train.boosting$error)
100 * (1 -  test.boosting$error)

100

72

Naive Bayes

score <- function(obs, exp) {
  return(100 * sum(obs == exp[,"C.Class"]) / length(obs))
}

nb <- naive_bayes(C.Class ~ ., data=train)
train.nb <- predict(nb, train)
 test.nb <- predict(nb, test)
score(train.nb, train)
score( test.nb, test)

81.4444444444444

75

simpel.go

out <- system2(c("./simpel", "data080.train", "data080.test"), stdout=TRUE, stderr=TRUE)
cat(out, sep="\n")

Training score:	 81.4%
Testing score:	 75.0%