def MoG_validation(K):
MoG_valid = mog.MoG("data2D.npy")
_, X_data, mu, _, sigma_2, log_pi, pi_np = MoG_valid.cluster(
K, D, B, 1.0 / 3.0)
# _, X_data, mu, _, sigma_2, log_pi, pi_np = MoG_valid.cluster(K, D, B)
loss_valid = MoG_valid.cal_loss(MoG_valid.validation.astype(np.float32),
mu, D, log_pi, sigma_2)
min_idx = MoG_valid.cal_min_idx(X_data, mu, np.sqrt(sigma_2), pi_np, D)
data = tf.ones(shape=[
B,
])
division = tf.unsorted_segment_sum(data, min_idx, K, name=None)
data_valid = tf.ones(shape=[
(B - (1 - 1 / 3) * B),
])
min_idx_valid = MoG_valid.cal_min_idx(
MoG_valid.validation.astype(np.float32), mu, np.sqrt(sigma_2), pi_np,
D)
division_valid = tf.unsorted_segment_sum(data_valid,
min_idx_valid,
K,
name=None)
with tf.Session():
print 'loss_validation:', loss_valid.eval()
print 'Total Proportion:', division.eval() / 10000
plot.plot_cluster(min_idx.eval(), X_data, mu, K)
plot.plot_valid_cluster(min_idx_valid.eval(), MoG_valid.validation, mu,
K)
def k_comparison(K):
D = 2
B = 10000
KM = km.k_mean("data2D.npy")
_, segment_ids, X_data, mu= KM.cluster(K, D, B)
data = tf.ones(shape = [B,])
division = tf.unsorted_segment_sum(data, segment_ids, K, name=None)
with tf.Session():
print "K =",K,":",division.eval()/10000
plot.plot_cluster(segment_ids, X_data, mu, K)
def k_comparison(K):
D = 2
B = 10000
KM = km.k_mean("data2D.npy")
_, segment_ids, X_data, mu= KM.cluster(K, D, B)
data = tf.ones(shape = [B,])
division = tf.unsorted_segment_sum(data, segment_ids, K, name=None)
with tf.Session():
print "K =",K,":",division.eval()/10000
plot.plot_cluster(segment_ids, X_data, mu, K)
def MoG_validation(K):
MoG_valid = mog.MoG("data2D.npy")
_, X_data, mu, _, sigma_2, log_pi, pi_np = MoG_valid.cluster(K, D, B, 1.0/3.0)
loss_valid = MoG_valid.cal_loss(MoG_valid.validation.astype(np.float32), mu, D, log_pi, sigma_2)
min_idx = MoG_valid.cal_min_idx(X_data, mu, np.sqrt(sigma_2), pi_np, D)
data = tf.ones(shape = [B,])
division = tf.unsorted_segment_sum(data, min_idx, K, name=None)
with tf.Session():
print 'loss_validation:', loss_valid.eval()
print 'Total Proportion:', division.eval()/10000
plot.plot_cluster(min_idx.eval(), X_data, mu, K)
else:
loss_prv = loss_train
if (epoch % 100 == 0):
print 'epoch', epoch
print 'loss', loss_train
print 'K =', K
print 'loss_training:', loss_train
print 'centroid:', mu_final
print 'pi_k:', tf.exp(pi_log).eval()
print 'sigma_square:', sigma_square
min_idx = self.cal_min_idx(self.train, mu_final, np.sqrt(sigma_square),
tf.exp(pi_log), D).eval()
return res_loss, X_data, mu_final, min_idx, sigma_square, pi_log, tf.exp(
pi_log).eval()
# When K = 3, compute the loss function for the whole data
K = 3
B = 10000
D = 2
mog = MoG("data2D.npy")
res_loss, X_plot, mu_plot, min_idx, _, _, pi_1 = mog.cluster(K, D, B)
plot.plot_loss(res_loss)
plot.plot_cluster(min_idx, X_plot, mu_plot, K)
min_idx = []
record, loss_prv = 0, 0
for epoch in range(training_epochs):
loss_np, min_idx, mu, _ = sess.run([loss, cluster, Y, train_op], feed_dict={X: X_train})
if record == 200:
break
elif loss_prv == loss_np:
record += 1
else:
loss_prv = loss_np
res_loss.append(loss_np)
'''
if(epoch % 200 == 0):
print epoch, ':' ,loss_np
# print Y.eval()
'''
return res_loss, min_idx, X_data, mu
K = 3 # Define 3 clusters
D = 2 #len(mean) # numbers of element per each dataset
B = 10000
km = k_mean("data2D.npy")
# Required argument: numbers of clusters, dimensions of points, numbers of points
res_loss, min_idx, X_data, mu= km.cluster(K, D, B, 1.0/3.0)
print mu
plot.plot_loss(res_loss)
plot.plot_cluster(min_idx, X_data, mu, K)
elif loss_prv == loss_train:
record += 1
else:
loss_prv = loss_train
if epoch % 100 == 0:
print "epoch", epoch
print "loss", loss_train
print "K =", K
print "loss_training:", loss_train
print "centroid:", mu_final
print "pi_k:", tf.exp(pi_log).eval()
print "sigma_square:", sigma_square
min_idx = self.cal_min_idx(self.train, mu_final, np.sqrt(sigma_square), tf.exp(pi_log), D).eval()
return res_loss, X_data, mu_final, min_idx, sigma_square, pi_log, tf.exp(pi_log).eval()
# When K = 3, compute the loss function for the whole data
K = 3
B = 10000
D = 2
mog = MoG("data2D.npy")
res_loss, X_plot, mu_plot, min_idx, _, _, pi_1 = mog.cluster(K, D, B)
plot.plot_loss(res_loss)
plot.plot_cluster(min_idx, X_plot, mu_plot, K)
record, loss_prv = 0, 0
for epoch in range(training_epochs):
loss_np, min_idx, mu, _ = sess.run([loss, cluster, Y, train_op],
feed_dict={X: X_train})
if record == 200:
break
elif loss_prv == loss_np:
record += 1
else:
loss_prv = loss_np
res_loss.append(loss_np)
'''
if(epoch % 200 == 0):
print epoch, ':' ,loss_np
# print Y.eval()
'''
return res_loss, min_idx, X_data, mu
K = 3 # Define 3 clusters
D = 2 #len(mean) # numbers of element per each dataset
B = 10000
km = k_mean("data2D.npy")
# Required argument: numbers of clusters, dimensions of points, numbers of points
res_loss, min_idx, X_data, mu = km.cluster(K, D, B, 1.0 / 3.0)
print mu
plot.plot_loss(res_loss)
plot.plot_cluster(min_idx, X_data, mu, K)