Tests on artificial data with 90% 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("data090.train", header=TRUE, sep="\t", quote="", row.names=1)
test  = read.table("data090.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
168	C	1.A2	2.B2	3.C1	4.B1	5.B1	6.C3	7.C2	8.B1	9.B1	⋯	11.C1	12.A1	13.A2	14.C4	15.B1	16.C4	17.B3	18.B1	19.A2	20.C3
391	A	1.B3	2.A1	3.A1	4.B2	5.A3	6.C2	7.B4	8.C1	9.B4	⋯	11.A1	12.C1	13.C6	14.A1	15.B1	16.A1	17.A2	18.C1	19.A2	20.B2
373	B	1.B2	2.B2	3.C2	4.B4	5.B1	6.A2	7.C2	8.B1	9.A1	⋯	11.A1	12.A1	13.C3	14.B1	15.B1	16.B1	17.A3	18.A1	19.B4	20.B3
488	C	1.C3	2.C6	3.A1	4.B4	5.B1	6.C3	7.B4	8.A1	9.C2	⋯	11.B2	12.B1	13.C3	14.C1	15.B1	16.A1	17.B3	18.A1	19.A2	20.A1
242	C	1.C2	2.B2	3.C2	4.B2	5.B1	6.C3	7.A1	8.C2	9.C3	⋯	11.C2	12.A1	13.B1	14.C1	15.A4	16.B1	17.B3	18.C1	19.B4	20.A1
781	C	1.B6	2.A1	3.C5	4.A3	5.A2	6.C1	7.B4	8.B1	9.A1	⋯	11.A1	12.C1	13.B2	14.A1	15.A4	16.C1	17.B5	18.A2	19.A1	20.A1
194	B	1.C2	2.A1	3.B2	4.A4	5.B1	6.A2	7.A2	8.B1	9.C3	⋯	11.A1	12.B1	13.B2	14.C1	15.B1	16.C3	17.B1	18.B1	19.B1	20.B1
893	B	1.A2	2.B2	3.A2	4.B2	5.A9	6.B2	7.B2	8.B1	9.C1	⋯	11.A1	12.A1	13.C3	14.A1	15.B1	16.B1	17.B1	18.C2	19.B4	20.B3
324	B	1.B6	2.C1	3.A2	4.B2	5.C3	6.B2	7.A2	8.B1	9.C3	⋯	11.A1	12.B1	13.B2	14.A1	15.B1	16.C4	17.B5	18.B1	19.A2	20.A1
430	B	1.C3	2.A1	3.A2	4.C4	5.A9	6.B1	7.A1	8.A8	9.B3	⋯	11.B2	12.A1	13.B1	14.C1	15.A2	16.B1	17.B3	18.A4	19.A2	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)

86.3333333333333

51

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

53

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)

65.6666666666667

50

simpel.go

out <- system2(c("./simpel", "data090.train", "data090.test"), stdout=TRUE, stderr=TRUE)
cat(out, sep="\n")

Training score:	 65.6%
Testing score:	 49.0%