# training
try:
myDML.train(moons_train)
except Exception as e:
print(e)
sys.exit(-1)
# evaluating
for K in [1, 3, 5]:
# testing with Euclidean_distance
predict_label = np.zeros(test_X.shape[0])
for i in range(test_X.shape[0]):
distance_vector = np.zeros(train_X.shape[0])
for j in range(train_X.shape[0]):
distance_vector[j] = myDML.Euclidean_distance(
test_X[i], train_X[j])
labels_of_K_neighbor = train_Y[distance_vector.argsort()[0:K]]
predict_label[i] = collections.Counter(
labels_of_K_neighbor).most_common(n=1)[0][0]
test_error = np.sum(predict_label != test_Y) / test_Y.shape[0]
print('Euclidean_distance+knn(k=' + str(K) + '): \t%f' % test_error)
# testing with your distance
predict_label = np.zeros(test_X.shape[0])
for i in range(test_X.shape[0]):
distance_vector = np.zeros(train_X.shape[0])
for j in range(train_X.shape[0]):
distance_vector[j] = myDML.distance(test_X[i], train_X[j])
labels_of_K_neighbor = train_Y[distance_vector.argsort()[0:K]]
predict_label[i] = collections.Counter(
labels_of_K_neighbor).most_common(n=1)[0][0]
# training
try:
myDML.train(moons_train)
except Exception as e:
print(e)
sys.exit(-1)
# evaluating
for K in [1, 3, 5]:
# testing with Euclidean_distance
predict_label = np.zeros(test_X.shape[0])
for i in range(test_X.shape[0]):
distance_vector = np.zeros(train_X.shape[0])
for j in range(train_X.shape[0]):
distance_vector[j] = myDML.Euclidean_distance(
test_X[i], train_X[j]) #计算测试集和训练集之间的距离
labels_of_K_neighbor = train_Y[distance_vector.argsort()
[0:K]] #取前k小个进行投票
predict_label[i] = collections.Counter(
labels_of_K_neighbor).most_common(
n=1)[0][0] #Counter计算元素的个数 most_common取TopN列表
test_error = np.sum(predict_label != test_Y) / test_Y.shape[0]
print('Euclidean_distance+knn(k=' + str(K) + '): \t%f' % test_error)
# testing with your distance
predict_label = np.zeros(test_X.shape[0])
for i in range(test_X.shape[0]):
distance_vector = np.zeros(train_X.shape[0])
for j in range(train_X.shape[0]):
distance_vector[j] = myDML.distance(test_X[i], train_X[j])
labels_of_K_neighbor = train_Y[distance_vector.argsort()[0:K]]