class1 = np.array(class1)
class2 = np.array(class2)
test_lines = [re.split("\s+", line) for line in lines[N:] if line]
test_data = np.array([line[:1] for line in test_lines], dtype=np.float)
test_labels = []
for line in test_lines:
y = line[-1]
y = float(y.rstrip())
test_labels.append(y)
return class1, class2, test_data, test_labels
if __name__ == '__main__':
c1, c2, tests, test_labels = read_data()
m = PGM()
# train
m.fit(c1, c2)
errors = 0
# predict
for i, t in enumerate(tests):
y = m.predict(t)
if y != test_labels[i]:
errors += 1
print '%s errors out of %s instance.' % (errors, len(tests))
print 'accuracy %s' % ((len(tests) - errors) * 1.0 / len(tests))
class2 = np.array([line[:4] for line in lines if line[-1] == 'Iris-versicolor'], dtype=np.float)
class3 = np.array([line[:4] for line in lines if line[-1] == 'Iris-virginica'], dtype=np.float)
test_lines = [line.split(",") for line in rawlines[N:] if line]
test_data = np.array([line[:4] for line in test_lines], dtype=np.float)
test_labels = []
for line in test_lines:
if line[-1] == 'Iris-setosa':
test_labels.append(0)
elif line[-1] == 'Iris-versicolor':
test_labels.append(1)
elif line[-1] == 'Iris-virginica':
test_labels.append(2)
return class1, class2, class3, test_data, test_labels
if __name__ == '__main__':
c1, c2, c3, tests, test_labels = read_data()
m = PGM()
# train
m.fit(c1, c2, c3)
errors = 0
# predict
for i, t in enumerate(tests):
y = m.predict(t)
if y != test_labels[i]:
errors += 1
print '%s errors out of %s instance.' %(errors, len(tests))
class2 = np.array(class2)
test_lines = [re.split("\s+", line) for line in lines[N:] if line]
test_data = np.array([line[:1] for line in test_lines], dtype=np.float)
test_labels = []
for line in test_lines:
y = line[-1]
y = float(y.rstrip())
test_labels.append(y)
return class1, class2, test_data, test_labels
if __name__ == "__main__":
c1, c2, tests, test_labels = read_data()
m = PGM()
# train
m.fit(c1, c2)
errors = 0
# predict
for i, t in enumerate(tests):
y = m.predict(t)
if y != test_labels[i]:
errors += 1
print "%s errors out of %s instance." % (errors, len(tests))
print "accuracy %s" % ((len(tests) - errors) * 1.0 / len(tests))
