采用caret包train函数进行随机森林参数寻优，代码如下，出现The tuning parameter grid should have columns mtry
上网找了很多回答，解释为随机森林可供寻优的参数只有mtry，但是一个一个更换ntree参数比较麻烦，请问只能用这种方法吗？
fit <- train(x=Csoc[,-c(1:5)],
y=Csoc[,5],
data = Csoc[,-c(1:4)],
method = "rf",
trControl = trainControl(method = "repeatedcv",
number = 10,repeats = 3),
tuneGrid = expand.grid(mtry = seq(3,21,3),
ntree = seq(300,700,100)),
#ntree = 300,
metric = "Rsquared")
plot(fit)
fit
