def main():
significant_posts = defaultdict(int)
for post in data.posts:
try:
if data.posts[post]["day"][1] - data.posts[post]["day"][
2] >= threshold:
significant_posts[data.posts[post]['title']] = data.posts[
post]["day"][1] - data.posts[post]["day"][2]
except KeyError:
continue
#print significant_posts[data.posts[post]['day']]
print len(significant_posts)
word_scores, weights = metrics()
#query = raw_input("Please enter the query that you would like to rate: ")
score = 0
for key, value in significant_posts.iteritems():
tokens = {}
token.get_tokens(key, tokens, 0, 1)
for t in tokens:
score += np.log10(word_scores[t] * weights[t])
print key
print "Post Score: " + str(value) + " Score: " + str(
score / len(tokens)) + "\n"
def evaluate(category_probabilities, post):
prob_per_category = defaultdict(float)
total_value = 0
for key, value in category_probabilities.iteritems():
total_value += len(value)
tokens = {}
token.get_tokens(data.posts[post]['title'], tokens, 0, 0)
for t in tokens:
for key, value in category_probabilities.iteritems():
if t in value:
prob_per_category[key] += math.log(value[t])
for key, value in prob_per_category.iteritems():
print len(category_probabilities[key]) / total_value
value *= len(category_probabilities[key]) / total_value
#DEBUG OUTPUT
if True:
print "Probability of true:\t" + str(prob_per_category["successful"])
print "Probability of lie:\t" + str(prob_per_category["fail"])
return min(prob_per_category.iteritems(), key=operator.itemgetter(1))[0]
def evaluate(category_probabilities, post):
prob_per_category = defaultdict(float)
total_value = 0
for key, value in category_probabilities.iteritems():
total_value += len(value)
tokens = {}
token.get_tokens(data.posts[post]['title'], tokens, 0, 0)
for t in tokens:
for key, value in category_probabilities.iteritems():
if t in value:
prob_per_category[key] += math.log(value[t])
for key, value in prob_per_category.iteritems():
print len(category_probabilities[key])/total_value
value *= len(category_probabilities[key])/total_value
#DEBUG OUTPUT
if True:
print "Probability of true:\t" + str(prob_per_category["successful"])
print "Probability of lie:\t" + str(prob_per_category["fail"])
return min(prob_per_category.iteritems(), key=operator.itemgetter(1))[0]
def vocabulary(posts):
vocab = defaultdict(int)
for post in posts:
tokens = {}
token.get_tokens(data.posts[post]['title'], tokens, 0, 1)
for t in tokens:
vocab[t] += 1
return vocab
def vocabulary(posts):
vocab = defaultdict(int)
for post in posts:
tokens = {}
token.get_tokens(data.posts[post]['title'], tokens, 0, 1)
for t in tokens:
vocab[t] += 1
return vocab
def main():
significant_posts = defaultdict(int)
for post in data.posts:
try:
if data.posts[post]["day"][1] - data.posts[post]["day"][2] >= threshold:
significant_posts[data.posts[post]['title']] = data.posts[post]["day"][1] - data.posts[post]["day"][2]
except KeyError:
continue
#print significant_posts[data.posts[post]['day']]
print len(significant_posts)
word_scores, weights = metrics()
#query = raw_input("Please enter the query that you would like to rate: ")
score = 0
for key, value in significant_posts.iteritems():
tokens = {}
token.get_tokens(key, tokens, 0, 1)
for t in tokens:
score += np.log10(word_scores[t] * weights[t])
print key
print "Post Score: " + str(value) + " Score: " + str(score/len(tokens)) + "\n"
def metrics():
word_scores = defaultdict(int)
num_words = defaultdict(int)
total_num_words = 0
highest_post = 0
title = ""
for post in data.posts:
tokens = {}
try:
#print post
#print data.posts[post]['title']
token.get_tokens(data.posts[post]['title'], tokens, 0, 1)
for t in tokens:
#Get upvotes
word_scores[t] += data.posts[post]['day'][1]
word_scores[t] -= data.posts[post]['day'][2]
num_words[t] += 1
total_num_words += 1
if data.posts[post]["day"][1] > highest_post:
title = post
highest_post = data.posts[post]["day"][1]
except KeyError:
continue
weights = defaultdict(float)
for key, value in num_words.iteritems():
#weights[key] = np.log10(total_num_words/float(value)+len(num_words));
weights[key] = 1 - (float(value) / total_num_words)
for key, value in word_scores.iteritems():
word_scores[key] = value / num_words[key]
print highest_post
print title
return word_scores, weights
def metrics():
word_scores = defaultdict(int)
num_words = defaultdict(int)
total_num_words = 0
highest_post = 0
title = ""
for post in data.posts:
tokens = {}
try:
#print post
#print data.posts[post]['title']
token.get_tokens(data.posts[post]['title'], tokens, 0, 1)
for t in tokens:
#Get upvotes
word_scores[t] += data.posts[post]['day'][1]
word_scores[t] -= data.posts[post]['day'][2]
num_words[t] += 1
total_num_words += 1
if data.posts[post]["day"][1] > highest_post:
title = post
highest_post = data.posts[post]["day"][1]
except KeyError:
continue
weights = defaultdict(float)
for key, value in num_words.iteritems():
#weights[key] = np.log10(total_num_words/float(value)+len(num_words));
weights[key] = 1 - (float(value)/total_num_words)
for key, value in word_scores.iteritems():
word_scores[key] = value / num_words[key]
print highest_post
print title
return word_scores, weights