Instruction to run

Part 2

1. Run train_parmas.py to train the data, transition and emition parameters will be stored in /params folder
2. Run Part2.py to generate dev.p2.out for all language

Part 3

1. Run train_params.py ( if you already did this in part 2, no need to do it again, since it uses the same parameters for predictoin)
2. Run Part3.py to generate dev.p3.out for all language

Part 4

1. Run train_params.py ( if you already did this in part 2 or 3, no need to do it again, since it uses the same parameters for predictoin)
2. Run Part3.py to generate dev.p4.out for all language

Part 5

Just run Part5_perceptron.py and test.p5.out will be generated for EN and ES

Ensure the parameter are:

ITER = 80

TOP_train = 1

TOP_predict = 1

CLEAN_DATA = False

For dev data, set TEST = False

For set data, set TEST = True

Detailed implementation can be found on https://github.com/hellozhanghao/ML_Project.git