def get_model(num_exm,
num_train,
lens,
block_len,
blocks=1,
anomaly_prob=0.15):
print(
'Generating {0} sequences, {1} for training, each with {2} anomaly probability.'
.format(num_exm, num_train, anomaly_prob))
cnt = 0
X = []
Y = []
label = []
lblcnt = co.matrix(0.0, (1, lens))
for i in range(num_exm):
(exm, lbl,
marker) = ToyData.get_2state_anom_seq(lens,
block_len,
anom_prob=anomaly_prob,
num_blocks=blocks)
cnt += lens
X.append(exm)
Y.append(lbl)
label.append(marker)
# some lbl statistics
if i < num_train:
lblcnt += lbl
X = normalize_sequence_data(X)
return (SOHMM(X[0:num_train],
Y[0:num_train]), SOHMM(X[num_train:],
Y[num_train:]), SOHMM(X, Y), label)
def get_model(num_exm, num_train, lens, block_len, blocks=1, anomaly_prob=0.15):
print('Generating {0} sequences, {1} for training, each with {2} anomaly probability.'.format(num_exm, num_train, anomaly_prob))
cnt = 0
X = []
Y = []
label = []
for i in range(num_exm):
(exm, lbl, marker) = ToyData.get_2state_anom_seq(lens, block_len, anom_prob=anomaly_prob, num_blocks=blocks)
cnt += lens
X.append(exm)
Y.append(lbl)
label.append(marker)
X = remove_mean(X,1)
return (SOHMM(X[0:num_train],Y[0:num_train]), SOHMM(X[num_train:],Y[num_train:]), SOHMM(X,Y), label)
def plot_icml_toy_seqs():
from toydata import ToyData
lens = 600
blocks = [1, 2, 5, 10, 20, 40, 60, 100]
data = []
lbl = []
for i in blocks:
(exm, label, marker) = ToyData.get_2state_anom_seq(lens,
120,
anom_prob=1.0,
num_blocks=i)
data.append(exm)
lbl.append(label)
plt.figure(1)
style = ['-', '--', '-', '.-', '--', '-', '-', '-', '-', '--']
marker = ['D', 'o', '^', 'D', 's', '^', 'D', 's', 's', 'o']
colors = ['r', 'b', 'm', 'm', 'm', 'c', 'c', 'c', 'g', 'g']
alphas = [1., 0.9, .8, .6, .4, .8, .6, .4, 1., .6]
widths = [2, 4, 1, 4, 8, 1, 4, 8, 1, 8]
ys = []
for i in [0, -1]:
ys.append(i * 10)
plt.plot(range(lens), np.array(data[i]).T + ys[-1], \
color='r', linewidth=2, alpha=0.6, marker='')
plt.plot(range(lens), np.array(lbl[i]).T + ys[-1], \
color='k', linewidth=2, alpha=0.8, marker='')
#plt.yscale('log')
plt.xticks([0, 300, 600], ['0', '300', '600'], fontsize=16)
plt.yticks(ys, ['0%', '100%'], fontsize=16)
plt.ylabel('Percentage of disorganization', fontsize=18)
plt.xlabel('Sequence position', fontsize=18)
plt.legend(['Noisy observations', 'True state sequence'],
fontsize=18,
fancybox=True,
framealpha=0.7)
plt.show()
