def train_and_validate_viterbi2(_inputFile, _outputFile, _devFile,
_devOutputFile, _validateFile):
"""
Create the Preprocessor object
Train using the SG, EN, CN, FR datasets
Generate the representer, vocabulary and states and feed it into an Emission object
"""
preprocessor = Preprocessor(_inputFile)
representer = preprocessor.get_representer()
vocabulary = preprocessor.get_vocabulary()
states = preprocessor.get_states()
listOfWords = getAllTokens(_devFile)
"""
Create the Emission and Transition objects
Validate using the dev datasets
Label the input sequence and output the file as dev.p3.out
"""
emission = Emission(representer, vocabulary, states, listOfWords)
transition = Transition2()
transition.compute_params(preprocessor)
label_viterbi(_devFile, _devOutputFile, emission, transition)
"""
Calculate Validation Error
"""
evaluate(_validateFile, _devOutputFile)
def train_and_validate_emission(_inputFile, _outputFile, _devFile,
_devOutputFile, _validateFile):
"""
Create the Preprocessor object
Train using the SG, EN, CN, FR datasets
Generate the representer, vocabulary and states and feed it into an Emission object
"""
preprocessor = Preprocessor(_inputFile)
representer = preprocessor.get_representer()
vocabulary = preprocessor.get_vocabulary()
states = preprocessor.get_states()
"""
Create the Emission Object
Validate using the dev datasets
Label the input sequence and output the file as dev.p2.out
"""
emission = Emission(representer, vocabulary, states)
emission.labelSequence(_devFile, _devOutputFile)
"""
Calculate Validation Error
"""
evaluate(_validateFile, _devOutputFile)