if args.algorithm != 'Maxent': train_kwargs['algorithm'] = args.algorithm trainf = MaxentClassifier.train train_kwargs['max_iter'] = args.max_iter train_kwargs['min_ll'] = args.min_ll train_kwargs['min_lldelta'] = args.min_lldelta train_kwargs['trace'] = args.trace else: trainf = NaiveBayesClassifier.train if args.multi and args.binary: if args.trace: print 'training a multi-binary %s classifier' % args.algorithm classifier = MultiBinaryClassifier.train(labels, train_feats, trainf, **train_kwargs) else: if args.trace: print 'training a %s classifier' % args.algorithm classifier = trainf(train_feats, **train_kwargs) ################ ## evaluation ## ################ if not args.no_eval: if not args.no_accuracy: print 'accuracy: %f' % accuracy(classifier, test_feats) if args.multi and args.binary and not args.no_masi_distance:
if args.multi and args.binary:
raise NotImplementedError(
"cross-fold is not supported for multi-binary classifiers")
scoring.cross_fold(train_feats,
trainf,
accuracy,
folds=args.cross_fold,
trace=args.trace,
metrics=not args.no_eval,
informative=args.show_most_informative)
sys.exit(0)
if args.multi and args.binary:
if args.trace:
print('training multi-binary %s classifier' % args.classifier)
classifier = MultiBinaryClassifier.train(labels, train_feats, trainf)
else:
classifier = trainf(train_feats)
################
## evaluation ##
################
if not args.no_eval:
if not args.no_accuracy:
try:
print('accuracy: %f' % accuracy(classifier, test_feats))
except ZeroDivisionError:
print('accuracy: 0')
if args.multi and args.binary and not args.no_masi_distance:
print('average masi distance: %f' %
##############
## training ##
##############
trainf = nltk_trainer.classification.args.make_classifier_builder(args)
if args.cross_fold:
if args.multi and args.binary:
raise NotImplementedError ("cross-fold is not supported for multi-binary classifiers")
scoring.cross_fold(train_feats, trainf, accuracy, folds=args.cross_fold,
trace=args.trace, metrics=not args.no_eval, informative=args.show_most_informative)
sys.exit(0)
if args.multi and args.binary:
if args.trace:
print 'training multi-binary %s classifier' % args.classifier
classifier = MultiBinaryClassifier.train(labels, train_feats, trainf)
else:
classifier = trainf(train_feats)
################
## evaluation ##
################
if not args.no_eval:
if not args.no_accuracy:
try:
print 'accuracy: %f' % accuracy(classifier, test_feats)
except ZeroDivisionError:
print 'accuracy: 0'
if args.multi and args.binary and not args.no_masi_distance:
print 'average masi distance: %f' % (scoring.avg_masi_distance(classifier, test_feats))