Chapter 6, exercise 9

## -*- coding: utf-8 -*-
import nltk
import random

wholeTrainingCorpus = nltk.corpus.ppattach.attachments('training')
nounAttachmentCorpus = [inst for inst in wholeTrainingCorpus
                        if inst.attachment == 'N']
random.shuffle(nounAttachmentCorpus)

def get_features(inst):
    features = {}
    features['noun1'] = inst.noun1
    # features['noun2'] = inst.noun2
    # features['verb'] = inst.verb
    return features

featureSets = [(get_features(inst), inst.prep)
               for inst in nounAttachmentCorpus]

cutOff = len(featureSets) / 4

trainSet, testSet = featureSets[:cutOff], featureSets[cutOff:]

naiveBayesClassifier = nltk.NaiveBayesClassifier.train(trainSet)
decisionTreeClassifier = nltk.DecisionTreeClassifier.train(trainSet)

print 'Accuracy with the naive Bayes classifier : {0:.2f}%'.format(nltk.classify.accuracy(naiveBayesClassifier, testSet))
print 'team', naiveBayesClassifier.classify({'noun1': 'team'}), 'researchers'
print
print 'Accuracy with the decision tree classifier : {0:.2f}%'.format(nltk.classify.accuracy(decisionTreeClassifier, testSet))
print 'team', decisionTreeClassifier.classify({'noun1': 'team'}), 'researchers'