def labeledSentiment():
os.chdir('validation-results')
accuracyMeasure = AccuracyMeasures('labeledSentiment')
os.chdir('..')
os.chdir('labeledSentiment')
for filename in os.listdir('.'):
if '.txt' in filename and not 'readme' in filename:
f = open(filename, 'r');
reader = csv.reader(f, delimiter='\t', quoting=csv.QUOTE_NONE)
for line in reader:
if line[1]:
se = SentimentEngine(line[0], 'simple')
sentimentSum = sum([sentence.getAggregatePolarizedSentenceSentiment() for sentence in se.getSentences()])
sentiment = 0
sentiment = 1 if sentimentSum > 0 else (0 if sentimentSum < 0 else -1)
if sentiment == -1:
# print("Skipping:", line[0])
accuracyMeasure.skipFile.write(line[0] + '\n')
else:
accuracyMeasure.itemCount += 1
# print(line[0], line[1], sentiment)
if int(line[1]) == 0 and sentiment == 0:
accuracyMeasure.trueNegative += 1
accuracyMeasure.trueNegativeFile.write(line[0] + '\n')
if int(line[1]) == 1 and sentiment == 0:
accuracyMeasure.falseNegative += 1
accuracyMeasure.falseNegativeFile.write(line[0] + '\n')
if int(line[1]) == 0 and sentiment == 1:
accuracyMeasure.falsePositive += 1
accuracyMeasure.falsePositiveFile.write(line[0] + '\n')
elif int(line[1]) == 1 and sentiment == 1:
accuracyMeasure.truePositive += 1
accuracyMeasure.truePositiveFile.write(line[0] + '\n')
else:
accuracyMeasure.invalidFile.write(line[0] + '\n')
f.close()
accuracyMeasure.accuracyFile.write((tabulate([['Actual Positive', str(accuracyMeasure.truePositive), str(accuracyMeasure.falseNegative)], ['Actual Negative', str(accuracyMeasure.falsePositive), str(accuracyMeasure.trueNegative)]], headers=['', 'Classified Positive', 'Classified Negative'])) + '\n')
precision = accuracyMeasure.truePositive / float(accuracyMeasure.truePositive + accuracyMeasure.falsePositive)
recall = accuracyMeasure.truePositive / float(accuracyMeasure.truePositive + accuracyMeasure.falseNegative)
accuracyMeasure.accuracyFile.write('numItems: ' + str(accuracyMeasure.itemCount) + '\n')
accuracyMeasure.accuracyFile.write('Precision: ' + str(precision) + '\n')
accuracyMeasure.accuracyFile.write('Recall: ' + str(recall) + '\n')
accuracyMeasure.accuracyFile.write('PF: ' + str(accuracyMeasure.falsePositive/ float(accuracyMeasure.falsePositive + accuracyMeasure.trueNegative)) + '\n')
accuracyMeasure.accuracyFile.write('F-Measure: ' + str(2.0 / ((1.0 / precision) + (1.0 / recall))) + '\n')
accuracyMeasure.truePositiveFile.close()
accuracyMeasure.trueNegativeFile.close()
accuracyMeasure.falsePositiveFile.close()
accuracyMeasure.falseNegativeFile.close()
accuracyMeasure.skipFile.close()
accuracyMeasure.invalidFile.close()
os.chdir('..')
def proCons():
os.chdir('validation-results')
accuracyMeasure = AccuracyMeasures('pro-cons')
os.chdir('..')
os.chdir('pro-cons')
consFile = open('IntegratedCons.txt', 'r', encoding = "ISO-8859-1")
cons = re.findall(r'<Cons>(.*?)</Cons>', consFile.read())
prosFile = open('IntegratedPros.txt', 'r', encoding = "ISO-8859-1")
pros = re.findall(r'<Pros>(.*?)</Pros>', prosFile.read())
for con in cons:
se = SentimentEngine(con, 'simple')
sentimentSum = sum([sentence.getAggregatePolarizedSentenceSentiment() for sentence in se.getSentences()])
sentiment = 0
sentiment = 1 if sentimentSum > 0 else (0 if sentimentSum < 0 else -1)
if sentiment == -1:
# print("Skipping:", con)
accuracyMeasure.skipFile.write(con + '\n')
else:
accuracyMeasure.itemCount += 1
if sentiment == 0:
accuracyMeasure.trueNegative += 1
accuracyMeasure.trueNegativeFile.write(con + '\n')
elif sentiment == 1:
accuracyMeasure.falsePositive += 1
accuracyMeasure.falsePositiveFile.write(con + '\n')
for pro in pros:
se = SentimentEngine(pro, 'simple')
sentimentSum = sum([sentence.getAggregatePolarizedSentenceSentiment() for sentence in se.getSentences()])
sentiment = 0
sentiment = 1 if sentimentSum > 0 else (0 if sentimentSum < 0 else -1)
if sentiment == -1:
# print("Skipping:", pro)
accuracyMeasure.skipFile.write(pro + '\n')
else:
accuracyMeasure.itemCount += 1
if sentiment == 0:
accuracyMeasure.falseNegative += 1
accuracyMeasure.falseNegativeFile.write(pro + '\n')
elif sentiment == 1:
accuracyMeasure.truePositive += 1
accuracyMeasure.truePositiveFile.write(pro + '\n')
accuracyMeasure.accuracyFile.write((tabulate([['Actual Positive', str(accuracyMeasure.truePositive), str(accuracyMeasure.falseNegative)], ['Actual Negative', str(accuracyMeasure.falsePositive), str(accuracyMeasure.trueNegative)]], headers=['', 'Classified Positive', 'Classified Negative'])) + '\n')
precision = accuracyMeasure.truePositive / float(accuracyMeasure.truePositive + accuracyMeasure.falsePositive)
recall = accuracyMeasure.truePositive / float(accuracyMeasure.truePositive + accuracyMeasure.falseNegative)
accuracyMeasure.accuracyFile.write('numItems: ' + str(accuracyMeasure.itemCount) + '\n')
accuracyMeasure.accuracyFile.write('Precision: ' + str(precision) + '\n')
accuracyMeasure.accuracyFile.write('Recall: ' + str(recall) + '\n')
accuracyMeasure.accuracyFile.write('PF: ' + str(accuracyMeasure.falsePositive/ float(accuracyMeasure.falsePositive + accuracyMeasure.trueNegative)) + '\n')
accuracyMeasure.accuracyFile.write('F-Measure: ' + str(2.0 / ((1.0 / precision) + (1.0 / recall))) + '\n')
accuracyMeasure.truePositiveFile.close()
accuracyMeasure.trueNegativeFile.close()
accuracyMeasure.falsePositiveFile.close()
accuracyMeasure.falseNegativeFile.close()
accuracyMeasure.skipFile.close()
accuracyMeasure.invalidFile.close()
prosFile.close()
consFile.close()
os.chdir('..')
def reviewPolarity():
os.chdir('validation-results')
accuracyMeasure = AccuracyMeasures('review_polarity')
os.chdir('..')
os.chdir('review_polarity/txt_sentoken')
os.chdir('neg')
for filename in os.listdir('.'):
if '.txt' in filename:
f = open(filename, 'r')
se = SentimentEngine(f.read(), 'simple')
sentimentSum = sum([sentence.getAggregatePolarizedSentenceSentiment() for sentence in se.getSentences()])
sentiment = 0
sentiment = 1 if sentimentSum > 0 else (0 if sentimentSum < 0 else -1)
if sentiment == -1:
pass
# print("Skipping:", con)
# accuracyMeasure.skipFile.write(con + '\n')
else:
accuracyMeasure.itemCount += 1
if sentiment == 0:
accuracyMeasure.trueNegative += 1
# accuracyMeasure.trueNegativeFile.write(con + '\n')
elif sentiment == 1:
accuracyMeasure.falsePositive += 1
# accuracyMeasure.falsePositiveFile.write(con + '\n')
f.close()
os.chdir('..')
os.chdir('pos')
for filename in os.listdir('.'):
if '.txt' in filename:
f = open(filename, 'r')
se = SentimentEngine(f.read(), 'simple')
sentimentSum = sum([sentence.getAggregatePolarizedSentenceSentiment() for sentence in se.getSentences()])
sentiment = 0
sentiment = 1 if sentimentSum > 0 else (0 if sentimentSum < 0 else -1)
if sentiment == -1:
pass
# print("Skipping:", pro)
# accuracyMeasure.skipFile.write(pro + '\n')
else:
accuracyMeasure.itemCount += 1
if sentiment == 0:
accuracyMeasure.falseNegative += 1
#accuracyMeasure.falseNegativeFile.write(pro + '\n')
elif sentiment == 1:
accuracyMeasure.truePositive += 1
#accuracyMeasure.truePositiveFile.write(pro + '\n')
f.close()
os.chdir('..')
accuracyMeasure.accuracyFile.write((tabulate([['Actual Positive', str(accuracyMeasure.truePositive), str(accuracyMeasure.falseNegative)], ['Actual Negative', str(accuracyMeasure.falsePositive), str(accuracyMeasure.trueNegative)]], headers=['', 'Classified Positive', 'Classified Negative'])) + '\n')
precision = accuracyMeasure.truePositive / float(accuracyMeasure.truePositive + accuracyMeasure.falsePositive)
recall = accuracyMeasure.truePositive / float(accuracyMeasure.truePositive + accuracyMeasure.falseNegative)
accuracyMeasure.accuracyFile.write('numItems: ' + str(accuracyMeasure.itemCount) + '\n')
accuracyMeasure.accuracyFile.write('Precision: ' + str(precision) + '\n')
accuracyMeasure.accuracyFile.write('Recall: ' + str(recall) + '\n')
accuracyMeasure.accuracyFile.write('PF: ' + str(accuracyMeasure.falsePositive/ float(accuracyMeasure.falsePositive + accuracyMeasure.trueNegative)) + '\n')
accuracyMeasure.accuracyFile.write('F-Measure: ' + str(2.0 / ((1.0 / precision) + (1.0 / recall))) + '\n')
accuracyMeasure.truePositiveFile.close()
accuracyMeasure.trueNegativeFile.close()
accuracyMeasure.falsePositiveFile.close()
accuracyMeasure.falseNegativeFile.close()
accuracyMeasure.skipFile.close()
accuracyMeasure.invalidFile.close()
os.chdir('../..')
