def run(name, t, cp=None, verbosity=1):
level = {0:logging.ERROR, 1:logging.WARNING, 2:logging.INFO, 3:logging.DEBUG}[verbosity]
logging.basicConfig(level=level)
# learn SPN
if cp is not None: knobs.cluster_penalty = cp
knobs.min_instances = 2
trn, vld, tst, schema = learn.load_data(name)
start = time.time()
net = gens.learn_spn(trn, schema, t)
#learn.smooth_network(net, vld, verbosity>0)
tst_llh = net.llh(tst)
vld_llh = net.llh(vld)
print name, '\tt: %.5f'%t, '\tcp: ', cp, '\ttime:%.1f'%(time.time()-start), '\ttree', len(net.pot), 'va:%8.4f'%vld_llh, 'te:%8.4f'%tst_llh
return net, vld_llh, tst_llh
"""
from matplotlib import pyplot as plt
import learn
import numpy as np
import math
import cost_function
def plot(x,y,xlabel,ylabel):
plt.plot(x,y,'x')
plt.xlabel(xlabel)
plt.ylabel(ylabel)
plt.show()
d = learn.load_data("data/train_four.csv")
chars_in_description = map(len,d.description)
chars_in_summary = map(len,d.summary)
num_votes = d.num_votes
num_views = d.num_views
num_comments = d.num_comments
latitude = d.latitude
longitude = d.longitude
log_num_views = map(math.log,num_views + 1)
log_mean_views = np.mean(log_num_views)
mean_views = np.exp(log_mean_views) - 1
print(mean_views)
num_points = len(d.num_views)
def plot_multiple():
