-
Notifications
You must be signed in to change notification settings - Fork 0
/
Classifier.py
72 lines (58 loc) · 3.12 KB
/
Classifier.py
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
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
from math import log10
from Evaluation import Eval
from TrainingModelFactory import TrainingModelFactory, BYOMTrainingModel
from VocabularyFactory import VocabularyFactory, CaseInsensitiveAlphabetChars, IsAlphaChars
class Classifier:
def __init__(self, training_file, test_file, byom, vocabulary=None, ngram_size=None, smoothing_value=None):
self.test_file = test_file
if byom:
self.vocabulary = IsAlphaChars()
self.trace_file = 'trace_myModel.txt'
self.eval_file = 'eval_myModel.txt'
self.training_model = BYOMTrainingModel(self.vocabulary, 1*10**-50, training_file)
else:
self.vocabulary = VocabularyFactory.get_vocabulary(vocabulary)
# given δ must be within [0 ... 1]
# if the given δ is smaller than 0, use 0
# if the given δ is larger than 1, use 1
smoothing_value = max(float(smoothing_value), 0)
if smoothing_value is not 0:
smoothing_value = min(float(smoothing_value), 1)
self.trace_file = 'trace_' + vocabulary + '_' + ngram_size + '_' + str(smoothing_value) + '.txt'
self.eval_file = 'eval_' + vocabulary + '_' + ngram_size + '_' + str(smoothing_value) + '.txt'
self.training_model = TrainingModelFactory.get_nb_training_model(self.vocabulary, ngram_size,
smoothing_value,
training_file)
def classify(self):
self.training_model.train()
self.test()
def get_most_frequent_ngrams(self):
self.training_model.train()
return self.training_model.get_ten_most_frequent_ngrams()
def test(self):
input_file = open(self.test_file, 'r', encoding="utf-8")
output_file = open(self.trace_file, 'w', encoding="utf-8")
for line in input_file:
# skip empty lines
if line is "\n":
continue
partitioned_line = line.split(maxsplit=3)
id = partitioned_line[0]
actual_language = partitioned_line[2]
tweet = partitioned_line[3]
if isinstance(self.vocabulary, CaseInsensitiveAlphabetChars):
tweet = tweet.lower()
highest_score = None
language_with_highest_score = None
for language in self.training_model.language_data.keys():
score = self.training_model.get_language_score_of_tweet(language, tweet)
if highest_score is None or highest_score < score:
highest_score = score
language_with_highest_score = language
languages_match = 'correct' if language_with_highest_score == actual_language else 'wrong'
output_file.write(str.join(' ', [id, language_with_highest_score, str(highest_score), actual_language,
languages_match]) + '\n')
input_file.close()
output_file.close()
eval = Eval(self.trace_file, self.eval_file)
eval.write_to_file()