Hi,
I have a question on using svm{e1071} for a classification task:
No matter how I split the data into training and test, I always end with a
perfect accuracy in training but sensitivity = 0 for test. One example is
like this
1 2
1 209 0
2 0 67
pred1
1 2
1 47 0
2 17 0
My question is, is there anything wrong with the following call:
m2 <- best.svm(class~., data=x1, gamma=2^(-3:3), cost=2^(0:5)) # x1 is
training data
pred1 <- predict(m2, x3) # x3 is test data
Thanks!
--
Weiwei Shi, Ph.D
Research Scientist
GeneGO, Inc.
"Did you always know?"
"No, I did not. But I believed..."
---Matrix III
[[alternative HTML version deleted]]