library(datasets)
data(iris)

ntrain <- 30
xc1 <- iris[1:50,1:4]
xc2 <- iris[51:100,1:4]

seqc1 <- sample(50)
xc1treina <- xc1[seqc1[1:ntrain],]
yc1treina <- matrix(0, nrow = ntrain)

seqc2 <- sample(50)
xc2treina <- xc2[seqc2[1:ntrain],]
yc2treina <- matrix(1, nrow = ntrain)

xc1teste <- xc1[seqc1[(ntrain+1):50],]
yc1teste <- matrix(0, nrow=(50 - ntrain))

xc2teste <- xc2[seqc2[(ntrain+1):50],]
yc2teste <- matrix(1, nrow = (50 - ntrain))

xin <- as.matrix(rbind(xc1treina, xc2treina))
yd <-  rbind(yc1treina, yc2treina)

xinteste <- as.matrix(rbind(xc1teste, xc2teste))
yteste <- rbind(yc1teste, yc2teste)

retlist <- trainperceptron(xin, yd, 0.1, 0.01, 100, 1)

wt <- as.matrix(unlist(retlist[1]))
yt <- yperceptron(xinteste, wt, 1)
erroteste <- yteste - yt

w <- train.perceptron(xin, yd, degrau)

yt <- degrau(func.perceptron(xinteste, w))
erroteste <- yteste - yt