def predict_svr(clf, data):
feature_vector = []
scores = []
predictions = []
fo = codecs.open(data, encoding='utf-8')
lines = fo.readlines()
fo.close()
client = MongoClient('localhost', 27017)
db = client['nlprokz']
glove = db.glove
line = 0
for each_line in lines:
row = each_line.split('\n')[0].split('\t')
essay_set = int(row[1])
scores.append(get_score(essay_set, row))
words = [i for i in each_line.lower().split()]
word_list = []
for word in words:
if not d.check(word):
suggest = d.suggest(word)
if (len(suggest) > 0):
word_list.append(suggest[0])
else:
word_list.append(word)
essay_vector = np.array([0.0 for i in xrange(300)])
for i in glove.find({"gram": {"$in": word_list}}):
word_vector = np.array([float(n) for n in i['glove_vector']])
if (len(word_vector) == 300):
essay_vector += word_vector
feature_vector.append(essay_vector / len(words))
predictions = clf.predict(np.array(feature_vector))
return get_average_kappa(np.array(scores), np.array(predictions))
def predict_svr(clf, data):
feature_vector = []
scores = []
predictions = []
fo = codecs.open(data, encoding='utf-8')
lines = fo.readlines()
fo.close()
line = 0
for each_line in lines:
row = each_line.split('\n')[0].split('\t')
vector = []
# Ignore the heading line
if line < 1:
line += 1
continue
if line % 50 == 0:
print('Validation sample: ' + str(line))
e = Essay(row, store_score=True)
f = e.features
for i in sorted(f.__dict__.keys()):
vector.append(f.__dict__[i])
scores.append(e.score)
feature_vector.append(vector)
line += 1
predictions = clf.predict(np.array(feature_vector))
return get_average_kappa(np.array(scores), np.array(predictions))
def predict_sklearn_random_forest(model, test_data_dump):
test_data = util.load_object(test_data_dump)
predictions = []
targets = []
targets = test_data[:, -1]
predictions = model.predict(test_data[:, :-1])
return get_average_kappa(targets, predictions)
def predict_random_forest(model, test_data_dump):
test_data = util.load_object(test_data_dump)
predictions = []
targets = []
for sample in test_data:
targets.append(sample[-1])
predictions.append(model.predict(sample[:-1]))
return get_average_kappa(targets, predictions)
def predict_svr(clf, data):
feature_vector = []
scores = []
predictions = []
fo = codecs.open(data, encoding='utf-8')
lines = fo.readlines()
fo.close()
client = MongoClient('localhost', 27017)
db = client['nlprokz']
glove = db.glove
line = 0
for each_line in lines:
row = each_line.split('\n')[0].split('\t')
essay_set = int(row[1])
scores.append(get_score(essay_set, row))
words = [i for i in each_line.lower().split()]
word_list = []
for word in words:
if not d.check(word):
suggest = d.suggest(word)
if (len(suggest) > 0):
word_list.append(suggest[0])
else:
word_list.append(word)
essay_vector = np.array([0.0 for i in xrange(300)])
for i in glove.find({"gram": {"$in": word_list}}):
word_vector = np.array([float(n) for n in i['glove_vector']])
if (len(word_vector) == 300):
essay_vector += word_vector
feature_vector.append(essay_vector / len(words))
# for each_line in lines:
# row = each_line.split('\n')[0].split('\t')
# vector = []
# # Ignore the heading line
# if line < 1:
# line += 1
# continue
# if line % 50 == 0:
# print('Validation sample: '+str(line))
# e = Essay(row, store_score = True)
# f = e.features
# for i in sorted(f.__dict__.keys()):
# vector.append(f.__dict__[i])
# scores.append(e.score)
# feature_vector.append(vector)
# line += 1
predictions = clf.predict(np.array(feature_vector))
return get_average_kappa(np.array(scores), np.array(predictions))
def predict_svr(clf, test_data_dump):
test_data = util.load_object(test_data_dump)
targets = test_data[:, -1]
predictions = clf.predict(test_data[:, :-1])
return get_average_kappa(targets, predictions)
i = 0
with tf.Session() as sess:
tf.initialize_all_variables().run()
while (True):
sess.run(train_op, feed_dict={x_: feature_vector, y_: scores})
if i % PRINT_STEP == 0:
c = sess.run(cost,
feed_dict={
x_: valid_feature_vector,
y_: valid_scores
})
print('training cost:', c)
costs.append(c)
if len(costs) > 3 and is_increasing(costs[-3:]) and count > 2:
break
elif len(costs) > 3 and is_increasing(costs[-3:]):
count += 1
print count
# else:
# count = 0
i += 1
response = sess.run(y, feed_dict={x_: test_feature_vector})
print get_average_kappa(response, test_scores)
# incorrect = sess.run(error,{data: data, target: target})
# incorrect = sess.run(error, feed_dict={x_: feature_vector, y_: scores})
# print feature_vector[feature_vector.shape[0] - 1]
# print sess.run(prediction,feed_dict={x_: feature_vector})
#print sess.run(prediction,{data: [[[1],[0],[0],[1],[1],[0],[1],[1],[1],[0],[1],[0],[0],[1],[1],[0],[1],[1],[1],[0]]]})
# print('Epoch {:2d} error {:3.1f}%'.format(i + 1, 100 * incorrect))
sess.close()