1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
|
# Natural Language Toolkit - NaiveBayes
# Capable of classifying the test or gold data using Naive Bayes algorithm
#
# Author: Sumukh Ghodke <sumukh dot ghodke at gmail dot com>
#
# URL: <http://www.nltk.org/>
# This software is distributed under GPL, for license information see LICENSE.TXT
from nltk_contrib.classifier import instances as ins, decisionstump as ds, Classifier
from nltk_contrib.classifier.exceptions import invaliddataerror as inv
class NaiveBayes(Classifier):
def __init__(self, training, attributes, klass):
Classifier.__init__(self, training, attributes, klass)
self.post_probs, self.class_freq_dist = None, None
def train(self):
Classifier.train(self)
self.post_probs = self.training.posterior_probablities(self.attributes, self.klass)
self.class_freq_dist = self.training.class_freq_dist()
for klass_value in self.klass:
self.class_freq_dist.inc(klass_value)#laplacian smoothing
def classify(self, instances):
for instance in instances:
instance.classified_klass = self.estimate_klass(instance)
def estimate_klass(self, instance):
estimates_using_prob = {}
for klass_value in self.klass:
class_conditional_probability = self.class_conditional_probability(instance, klass_value)
estimates_using_prob[class_conditional_probability] = klass_value
keys = estimates_using_prob.keys()
keys.sort()#find the one with max conditional prob
return estimates_using_prob[keys[-1]]
def prior_probability(self, klass_value):
return self.class_freq_dist.freq(klass_value)
def posterior_probability(self, attribute, attribute_value, klass_value):
return self.post_probs.value(attribute, attribute_value, klass_value)
def class_conditional_probability(self, instance, klass_value):
class_cond_prob = 1.0
for attribute in self.attributes:
attr_value = instance.value(attribute)
class_cond_prob *= self.posterior_probability(attribute, attr_value, klass_value)
class_cond_prob *= self.prior_probability(klass_value)
return class_cond_prob
@classmethod
def can_handle_continuous_attributes(self):
return True
def is_trained(self):
return self.post_probs is not None and self.class_freq_dist is not None
|
