R exercise

Logistic Regression 10

Sign in to hand in your solution and get feedback.

train <- (Year < 2005) glm.fit <- glm(Direction ~ Lag1 + Lag2, data = Smarket, family = binomial, subset = train) predict(...)