def score_essay(essay_id):
from app import db, cache
essay = Essay.query.filter_by(id=essay_id).first()
manager = Manager(essay.question)
score = manager.score_essay(essay)
essay.predicted_score = score
essay.model = manager.get_latest_model()
db.session.commit()
try:
if math.isnan(e): essay[itr] = ""
except:
continue
score = list(tab1['score_swathi'])
threshold = 10
mean_error = []
std_error = []
for i in range(20):
print i, "th iteration"
train_essay, train_score, test_essay, test_score = divide_test_and_train(
essay, score, threshold)
print "samples for training is ", len(train_essay)
print "sample for testing is ", len(test_essay)
id = 'a'
manager = Manager(essay)
model_loc = manager.create_model(id,
text=train_essay,
scores=train_score,
MODEL_PATH=settings.MODEL_PATH)
print 'model is created and stored in ', model_loc
predicted_score = [
manager.score_essay(text=te, MODEL_PATH=model_loc)
for itera, te in enumerate(test_essay)
]
diff_score = [
abs(sc - predicted_score[i][0]) / 10.0
for i, sc in enumerate(test_score)
]
mean_error.append(np.mean(diff_score))
std_error.append(np.std(diff_score))
def create_model(question_id):
from app import cache
question = Question.query.filter_by(id=question_id).first()
manager = Manager(question)
manager.create_model()
cache.delete('{0}_question_status')
test_scores.append(temp_scores[itr])
return train_text, train_scores, test_text, test_scores
if __name__ == "__main__":
input_file = "/Users/alok/Documents/Data/Text_Mining/AES/The kings school_que_a_sec_abc_g123.csv"
tab1= pd.read_csv(input_file,header=0)
essay = list(tab1['text'])
for itr, e in enumerate(essay):
try:
if math.isnan(e): essay[itr] = ""
except:
continue
score = list(tab1['score_swathi'])
threshold = 10
mean_error = []
std_error = []
for i in range(20):
print i,"th iteration"
train_essay, train_score, test_essay,test_score = divide_test_and_train(essay,score, threshold)
print "samples for training is ", len(train_essay)
print "sample for testing is ", len(test_essay)
id ='a'
manager = Manager(essay)
model_loc = manager.create_model(id, text=train_essay, scores=train_score, MODEL_PATH = settings.MODEL_PATH)
print 'model is created and stored in ', model_loc
predicted_score=[manager.score_essay(text = te, MODEL_PATH = model_loc) for itera, te in enumerate(test_essay)]
diff_score = [abs(sc - predicted_score[i][0])/10.0 for i,sc in enumerate(test_score)]
mean_error.append(np.mean(diff_score))
std_error.append(np.std(diff_score))
print 'average of error is ', np.mean(mean_error), mean_error
print 'standard deviation of error is ',np.mean(std_error), std_error
