Feature selection and Model Tuning in R

@gupta24789 wrote:

Hello everyone,

Th following modifications are done to prepare the data ;
1. combine the train and test set data
2. missing value treatment
3. feature scaling i.e combine levels within categorical variable,make new variables for numerical predictors
4. feature extraction
5. tune the model for lm, glm, rpart, rf

But still my RMSE is 1150 on the test data set. Anyone can suggest me what I do next ?

Thanks !!!

Posts: 1

Participants: 1

Read full topic