def getFeature(item):
if item is None:
return
yr = item['year']
vec = fe.get(item['sentences1'],item['sentences2'])
titles = ([item['title1'],item['title2']])
# print [vec,titles,yr]
return ([vec,titles,yr,item['date1'],item['date2']])
def getFeature(item):
if item is None:
return
yr = item['year']
vec = fe.get(item['sentences1'], item['sentences2'])
titles = ([item['title1'], item['title2']])
# print [vec,titles,yr]
return ([vec, titles, yr, item['date1'], item['date2']])
def train(doubleSets):
bools = []
features = []
for item in doubleSets:
bools.append(item['year'])
vec = fe.get(item['sentences1'],item['sentences2'])
features.append(vec)
print "Training The Classifier."
clf.fit(features,bools)
def test(doubleSets):
bools = []
features = []
correct = 0
incorrect = 0
for item in doubleSets:
bools.append(item['year'])
vec = fe.get(item['sentences1'],item['sentences2'])
titles.append([item['title1'],item['title2']])
features.append(vec)
for feature in range(len(features)):
predict = clf.predict(np.array9[features[feature]]))
prob = clf.predict_proba(np.arrat([features[feature]]))
probs.append([predict,prob, bools[feature]])
def getFeature(item):
yr = item['year']
vec = fe.get(item['sentences1'], item['sentences2'])
titles = ([item['title1'], item['title2']])
return ([vec, titles, yr])
def getFeature(item):
yr = item['year']
vec = fe.get(item['sentences1'],item['sentences2'])
titles = ([item['title1'],item['title2']])
return ([vec,titles,yr])