qed qed - 1 month ago 23
R Question

Gamma changed in SVM with radial kernel

Here is the code:

library(e1071)
set.seed(1)
x = matrix(rnorm(200 * 2), ncol = 2)
x[1:100, ] = x[1:100, ] + 2
x[101:150, ] = x[101:150, ] - 2
y = c(rep(1, 150), rep(2, 50))
dat = data.frame(x = x, y = as.factor(y))
head(dat)
plot(x, col = y)
train = sample(200, 100)
svmfit = svm(y ~ ., data = dat[train, ], kernel = "radial", gammma = 1, cost = 1)
plot(svmfit, dat[train, ])
summary(svmfit)


The gamma parameter in the summary is different from what I set:

Call:
svm(formula = y ~ ., data = dat[train, ], kernel = "radial", gammma = 1, cost = 1)


Parameters:
SVM-Type: C-classification
SVM-Kernel: radial
cost: 1
gamma: 0.5

Number of Support Vectors: 36

( 18 18 )


Number of Classes: 2

Levels:
1 2


What went wrong?

Answer

You simply misspelled parameter, look closely gammma, how many ms do you see? It took me a while to see that there are more than 2.

Try

library(e1071)
set.seed(1)
x = matrix(rnorm(200 * 2), ncol = 2)
x[1:100, ] = x[1:100, ] + 2 
x[101:150, ] = x[101:150, ] - 2 
y = c(rep(1, 150), rep(2, 50))
dat = data.frame(x = x, y = as.factor(y))
head(dat)
plot(x, col = y)
train = sample(200, 100)
svmfit = svm(y ~ ., data = dat[train, ], kernel = "radial", gamma = 1, cost = 1)
plot(svmfit, dat[train, ])
summary(svmfit)

Shame on e1071 authors for not rising exception for specifying incorrect parameters though...

Comments