def main(args):
conn, cur = metrics.connect("cooking", "Ben-han")
user_ids = search_utilities.get_experts() + search_utilities.get_nonexperts()
shuffle(user_ids)
data, labels = training_examples(cur, conn, user_ids)
train_data = data[:len(data)/2]
train_labels = labels[:len(labels)/2]
test_data = data[len(data)/2:]
test_labels = labels[len(labels)/2:]
ml.logistic_test(train_data, train_labels, train_data, train_labels, cv=True)
def sample(cur, sample_size=50):
"""Samples num users out of the non-expert population that has asked or answered at least one question (active)"""
experts = tuple(search_utilities.get_experts())
query1 = """SELECT u.id
FROM se_user u
INNER JOIN post p
ON p.owner_user_id = u.id
WHERE u.id <> -1
AND u.id NOT IN %(experts)s
AND p.post_type_id = 2
GROUP BY u.id HAVING Count(*) > 1
ORDER BY random()
LIMIT %(limit)s;"""
cur.execute(query1, {"experts": experts, "limit": sample_size})
return list(result[0] for result in cur)
def sample(cur, sample_size = 50):
"""Samples num users out of the non-expert population that has asked or answered at least one question (active)"""
experts = tuple(search_utilities.get_experts())
query1 = """SELECT u.id
FROM se_user u
INNER JOIN post p
ON p.owner_user_id = u.id
WHERE u.id <> -1
AND u.id NOT IN %(experts)s
AND p.post_type_id = 2
GROUP BY u.id HAVING Count(*) > 1
ORDER BY random()
LIMIT %(limit)s;"""
cur.execute(query1, {"experts": experts, "limit": sample_size})
return list(result[0] for result in cur)
def main(args):
conn, cur = connect("cooking", "Ben-han")
experts = search_utilities.get_experts()
nonexperts = search_utilities.get_nonexperts()
# plot_individual_elo(cur, conn, 95, 'red')
# plt.savefig("output/codegolf.png")
# plt.show()
# Expert plots
# print "Plotting average expert Auth"
# plot_avg_auth(cur, experts, 'blue')
print "Plotting average expert PageRank"
plot_avg_pagerank(cur, experts, 'green', label = "Expert")
# print "Plotting average expert ELO"
# plot_avg_elo(cur, conn, experts, 'blue')
# print "Plotting average expert Cau"
# plot_avg_cau(cur, conn, experts, 'purple')
# plt.legend(loc = 4)
# plt.savefig("output/expert_combined.png")
# plt.show()
# Non-expert plots
# print "Plotting average nonexpert Auth"
# plot_avg_auth(cur, nonexperts, 'blue')
print "Plotting average nonexpert PageRank"
plot_avg_pagerank(cur, nonexperts, 'blue', label = "Non-expert")
# print "Plotting average nonexpert ELO"
# plot_avg_elo(cur, conn, nonexperts, 'red')
# print "Plotting average nonexpert Cau"
# plot_avg_cau(cur, conn, nonexperts, 'green')
# plt.legend(loc = 4)
#plot_reputation_distribution(cur, 'green')
plt.suptitle("Avg. expert PageRank vs non-expert PageRank")
plt.xlabel("Lifetime percentage")
plt.ylabel("Rating")
plt.legend(loc = 4)
plt.savefig("output/expert_vs_nonexpert_pagerank.png")
plt.show()
def main(args):
conn, cur = connect("cooking", "Ben-han")
experts = search_utilities.get_experts()
nonexperts = search_utilities.get_nonexperts()
# plot_individual_elo(cur, conn, 95, 'red')
# plt.savefig("output/codegolf.png")
# plt.show()
# Expert plots
# print "Plotting average expert Auth"
# plot_avg_auth(cur, experts, 'blue')
print "Plotting average expert PageRank"
plot_avg_pagerank(cur, experts, 'green', label="Expert")
# print "Plotting average expert ELO"
# plot_avg_elo(cur, conn, experts, 'blue')
# print "Plotting average expert Cau"
# plot_avg_cau(cur, conn, experts, 'purple')
# plt.legend(loc = 4)
# plt.savefig("output/expert_combined.png")
# plt.show()
# Non-expert plots
# print "Plotting average nonexpert Auth"
# plot_avg_auth(cur, nonexperts, 'blue')
print "Plotting average nonexpert PageRank"
plot_avg_pagerank(cur, nonexperts, 'blue', label="Non-expert")
# print "Plotting average nonexpert ELO"
# plot_avg_elo(cur, conn, nonexperts, 'red')
# print "Plotting average nonexpert Cau"
# plot_avg_cau(cur, conn, nonexperts, 'green')
# plt.legend(loc = 4)
#plot_reputation_distribution(cur, 'green')
plt.suptitle("Avg. expert PageRank vs non-expert PageRank")
plt.xlabel("Lifetime percentage")
plt.ylabel("Rating")
plt.legend(loc=4)
plt.savefig("output/expert_vs_nonexpert_pagerank.png")
plt.show()
def is_expert(user_id):
"""
Returns true if the user is an expert.
"""
return user_id in search_utilities.get_experts()
