def extract_features_from_images(self):
if self.train_model:
num_samples = 10000
else:
num_samples = 10
# Take random samples from the images to train the model
random_idxs = np.random.randint(0, len(self.cars), num_samples)
test_cars = np.array(self.cars)[random_idxs]
test_notcars = np.array(self.notcars)[random_idxs]
print("Extracting Car Features")
fE_cars = featureExtractor()
car_features = fE_cars.get_features_from_images(
test_cars, self.size, self.nbins, self.orientation,
self.pixels_per_cell, self.cell_per_block)
print("Extracting Non Car Features")
fE_notcars = featureExtractor()
not_car_features = fE_notcars.get_features_from_images(
test_notcars, self.size, self.nbins, self.orientation,
self.pixels_per_cell, self.cell_per_block)
return car_features, not_car_features
def getReviewSentiment(app, tknRevs, classifier):
revAggSentiment = 0
for revList in tknRevs:
sentAggSentiment = 0
for sent in revList:
sent = unicode(sent.strip())
# print sent
featdata = extractor.featureExtractor(sent)
# pprint(featdata)
#pdb.set_trace()
cl = classifier.classify(featdata)
if cl == 'pos':
label = 1
elif cl == 'neutral':
label = 0
else:
label = -1
sentAggSentiment += label
revAggSentiment += sentAggSentiment
name = app['name'].encode('utf-8')
print "App: \t %s, Aggregate Review Sentiment: \t %s" % (name,
revAggSentiment)
return revAggSentiment
def getReviewSentiment(app, tknRevs, classifier):
revAggSentiment = 0
for revList in tknRevs:
sentAggSentiment = 0
for sent in revList:
sent = unicode(sent.strip())
# print sent
featdata = extractor.featureExtractor(sent)
# pprint(featdata)
#pdb.set_trace()
cl= classifier.classify(featdata)
if cl == 'pos':
label = 1
elif cl == 'neutral':
label = 0
else:
label = -1
sentAggSentiment += label
revAggSentiment += sentAggSentiment
name = app['name'].encode('utf-8')
print "App: \t %s, Aggregate Review Sentiment: \t %s" % (name , revAggSentiment)
return revAggSentiment
def getReviewSentiment(tknRevs, classifier):
revAggSentiment = 0
for sent in tknRevs:
sent = unicode(sent.strip())
featdata = extractor.featureExtractor(sent)
cl = classifier.classify(featdata)
if cl == 'pos':
label = 1
elif cl == 'neutral':
label = 0
else:
label = -1
revAggSentiment += label
return revAggSentiment
def getReviewSentiment(tknRevs, classifier):
revAggSentiment = 0
for sent in tknRevs:
sent = unicode(sent.strip())
featdata = extractor.featureExtractor(sent)
cl= classifier.classify(featdata)
if cl == 'pos':
label = 1
elif cl == 'neutral':
label = 0
else:
label = -1
revAggSentiment += label
return revAggSentiment
