Write a loop in Python that would be able to set train_cols equal to all possible combinations of features. The purpose of this is to be able to run a confusion matrix on each combination of train_cols and save the train_cols with the best value later on.

train_cols = ["feature1", "feature2", "feature3", "feature4", "feature5", "feature6", "feature7", "feature8", "feature9", "feature10"]

For example, if the contents of train_cols was ["a", "b", "c"] the correct code would set train_cols equal to the following possibilities during the loop (order does not matter):

train_cols = ["a"]

train_cols = ["b"]

train_cols = ["c"]

train_cols = ["a", "b"]

train_cols = ["a", "c"]

train_cols = ["b", "c"]

train_cols = ["a", "b", "c"]