本文整理匯總了Python中preprocessor.Preprocessor.ngram_tokenizer方法的典型用法代碼示例。如果您正苦於以下問題：Python Preprocessor.ngram_tokenizer方法的具體用法？Python Preprocessor.ngram_tokenizer怎麽用？Python Preprocessor.ngram_tokenizer使用的例子？那麽, 這裏精選的方法代碼示例或許可以為您提供幫助。您也可以進一步了解該方法所在類preprocessor.Preprocessor的用法示例。
在下文中一共展示了Preprocessor.ngram_tokenizer方法的1個代碼示例，這些例子默認根據受歡迎程度排序。您可以為喜歡或者感覺有用的代碼點讚，您的評價將有助於係統推薦出更棒的Python代碼示例。
示例1: __init__
# 需要導入模塊: from preprocessor import Preprocessor [as 別名]
# 或者: from preprocessor.Preprocessor import ngram_tokenizer [as 別名]
class MutualInformation:
def __init__(self, files_path='', classes={}, out_file='output.csv'):
# self.mi_terms looks like this {'term1': {'d': 3, 't': 4, 'mi': 2, },}
self.mi_terms = {}
# self.mi_classes looks like this {'d': 3, 't': 4}
self.mi_classes = {}
self.total_terms_count = 0
# Some configuration
self.files_path = files_path
self.out_file = out_file
self.classes = classes
self.files_prefixes = classes.keys()
self.class_names = [classes[prefix] for prefix in classes]
# For tokenizing, stemming, etc.
self.prep = Preprocessor(pattern='\W+', lower=True, stem=False, stemmer_name='porter', pos=False, ngram=1)
# Read terms from files, and fill self.mi_terms & self.mi_classes
def load_terms(self):
files = os.listdir(self.files_path)
for filename in files:
#print filename
terms = []
file_prefix = ''
if filename.startswith(self.files_prefixes[0]):
file_prefix = self.files_prefixes[0]
elif filename.startswith(self.files_prefixes[1]):
file_prefix = self.files_prefixes[1]
else:
continue
fd = open('%s/%s' % (self.files_path, filename), 'r')
file_data = fd.read()
fd.close()
terms = self.prep.ngram_tokenizer(text=file_data)
for term in terms:
self.total_terms_count += 1
if not self.mi_terms.has_key(term):
self.mi_terms[term] = {self.files_prefixes[0]: 0, self.files_prefixes[1]: 0}
self.mi_terms[term][file_prefix] += 1
if self.mi_classes.has_key(file_prefix):
self.mi_classes[file_prefix] += 1
else:
self.mi_classes[file_prefix] = 0
print self.mi_classes
# Term probablility
def pr_term(self, term):
term_count = self.mi_terms[term][self.files_prefixes[0]] + self.mi_terms[term][self.files_prefixes[1]]
total_count = self.mi_classes[self.files_prefixes[0]] + self.mi_classes[self.files_prefixes[1]]
return term_count * 1.00 / total_count
# Class probability
def pr_class(self, class_prefix):
class_count = self.mi_classes[class_prefix]
total_count = self.mi_classes[self.files_prefixes[0]] + self.mi_classes[self.files_prefixes[1]]
return class_count * 1.00 / total_count
# Posterior Probability Pr(term/class)
def pr_post(self, term, class_prefix):
term_count = self.mi_terms[term][class_prefix]
total_count = self.mi_classes[class_prefix]
return term_count * 1.00 / total_count
# Joint Probability Pr(term, class)
def pr_joint(self, term, class_prefix):
return self.pr_post(term, class_prefix) * self.pr_class(class_prefix)
# Q = 1- P
def q(self, p):
return (1 - p)
# Calculate Mutual Information
def calculate_mi(self):
for term in self.mi_terms:
mi = 0.0
for class_prefix in self.files_prefixes:
try:
mi += self.pr_joint(term, class_prefix) * math.log10(self.pr_post(term, class_prefix) / self.pr_term(term))
mi += self.q(self.pr_joint(term, class_prefix)) * math.log10(self.q(self.pr_post(term, class_prefix)) / self.q(self.pr_term(term)))
except:
# Ok, log(0), let's set mi = -1
mi = 0
self.mi_terms[term]['mi'] = mi
# Dump results into a CSV File
def mi2csv(self):
fd = open(self.out_file , 'w')
header_line = "Term, %s, %s, Mutual Info\n" % (self.classes[self.files_prefixes[0]], self.classes[self.files_prefixes[1]])
fd.write(header_line)
for term in self.mi_terms:
new_line = '%s, %f, %f, %f\n' % (term, self.mi_terms[term][self.files_prefixes[0]],
self.mi_terms[term][self.files_prefixes[1]], self.mi_terms[term]['mi'])
fd.write(new_line)
fd.close()