gistfile1.txt

amazon_test$ACTION1 = 1

#this is really just a placeholder variable so that the training examples wont be removed when I call match. 
#also, make sure you put this in the same column as it is in the training set...



amazon_test = sparse.model.matrix(~. - 1, data = amazon_test)


#turn the test set into a sparse matrix. the training is in this format already.



amazon_train = amazon_train[, c(which(complete.cases(match(colnames(amazon_train), colnames(amazon_test))) == TRUE))]



#to break down what this does, match takes the second argument and returns the numerical position of where it
#exists in the first argument.  If it doesn't find it, it returns NA.  so I'm telling it to look through a
#character vector of column names in the test file, relate them to what columns they correspond to in
#training, and only keep those.  The complete.cases wrapper removes all the NAs from the vector and gives
#me a nice numerical list.




amazon_test = amazon_test[, c(which(complete.cases(match(colnames(amazon_test), colnames(amazon_train))) == TRUE))]



#not only that, you need to run it again to take away the dummies that are in the test set but not in training



dim(amazon_train) ; dim(amazon_test)

#should have the same number of columns now