To train the program, 4 arguments must be provided. The first argument of the program always pertains to either training or predicting. If it is “train”, then the second argument must be the training file the user wants to use. The training file contains a list of sentences, with the indication whether they are English or Dutch at the start. The third argument specifies the hypothesis file name that the program should write out. It is the model that is to be used for the prediction. The fourth and final argument is the learning type, which could be Decision Tree (dt), or AdaBoost (ada).

train <file_to_train_from> <output_model_name>

To predict the languages of a file, “predict” should be the first argument. The second should be the name of the hypothesis file, or model, that the training algorithm wrote in the training step. The third and final argument should be the name of the test file for which the languages the user wishes to be identified. The main class will determine if the model is of type Decision Tree, or AdaBoost, and then call the method that predicts the language to its appropriate instance type.

predict <model_name>