Submit the prediction results in file result.txt. In result.txt, output the predicted authors of the emails in both the validation set and the test set, with each line containing one author. For example:
beck-s
beck-s
farmer-d
......
Output the predictions for validation set first and then the predictions for the test set.
You can make more than one submissions, but you should select only one as your final submission.
Before the competition is complete, you can only see the evaluation results on the validation set.
At the end, your system will be evaluated on the test set and produce a final accuracy score
(it might be different from the score obtained from the validation set).