def test_distance(self):
point1 = (20, 30, 40)
point2 = (20, 30, 40)
point3 = (250, 240, 230)
print "Running test 1.a"
assert kmeans.distance(point1, point2) == 0
print "Running test 1.b"
assert kmeans.distance(point1, point3) == 133100
def nearest(point, cluster_centers):
min_dist = FLOAT_MAX
m = np.shape(cluster_centers)[0] # 当前已经初始化的聚类中心的个数
for i in range(m):
# 计算point与每个聚类中心之间的距离
d = distance(point, cluster_centers[
i, ]) #计算当前点与已经初始化的每个点的距离，并且得到最小值，即假设该点与初始化的第二个簇中心最短
# 选择最短距离
if min_dist > d:
min_dist = d
return min_dist
def nearest(point, cluster_centers):
'''
计算point和cluster_centers之间的最小距离
input: point(mat):当前的样本点
cluster_centers(mat):当前已经初始化的聚类中心
output: min_dist(float):点point和当前的聚类中心之间的最短距离
'''
min_dist = FLOAT_MAX
m = np.shape(cluster_centers)[0] # 当前已经初始化的聚类中心的个数
for i in range(m):
# 计算point与每个聚类中心之间的距离
d = distance(point, cluster_centers[i, ])
# 选择最短距离
if min_dist > d:
min_dist = d
return min_dist
def generate_path(probability_matrix, state_prototypes, action_prototypes, state_indices, iterations):
state_index = random.sample(state_indices, 1)[0] # select random point
path = []
current_position = state_prototypes[state_index] # choose closest state prototype
for i in range(iterations):
path.append(current_position)
# following should not be the case but who knows what is going on with 2nd map
if np.sum(probability_matrix[state_index]) == 0:
best_action_index = np.argmax((probability_matrix[state_index]))
else:
# choose best action using matrix of joint probabilities
best_action_index = np.argmax((probability_matrix[state_index] / np.sum(probability_matrix[state_index])))
best_action = action_prototypes[best_action_index]
current_position = current_position + best_action
dist = [kmeans.distance(current_position, k) for k in state_prototypes]
min_index, = np.where(dist == np.min(dist)) # choose closest prototype
state_index = min_index[0]
return path
def problem2_4_4():
data, featureNames = load_data()
k = 6
iniCenters = orderedCenters(data, k)
clusters, centers = kmeans(data, k,
initialCenters=iniCenters)
avgAll = avg(data.keys(), data)
print "average: %s" % avgAll
print "########################"
print "cluster # -> (distance from average)"
print " [# in cluster]:[items in cluster]"
print ""
for c in clusters:
print "cluster %d -> %s: " % (c, distance(avgAll, centers[c]))
print " [%d]:%s " % (len(clusters[c]), clusters[c])
print ""
def generate_path(probability_matrix, state_prototypes, action_prototypes,
state_indices, iterations):
state_index = random.sample(state_indices, 1)[0] # select random point
path = []
current_position = state_prototypes[
state_index] # choose closest state prototype
for i in range(iterations):
path.append(current_position)
# following should not be the case but who knows what is going on with 2nd map
if np.sum(probability_matrix[state_index]) == 0:
best_action_index = np.argmax((probability_matrix[state_index]))
else:
# choose best action using matrix of joint probabilities
best_action_index = np.argmax(
(probability_matrix[state_index] /
np.sum(probability_matrix[state_index])))
best_action = action_prototypes[best_action_index]
current_position = current_position + best_action
dist = [kmeans.distance(current_position, k) for k in state_prototypes]
min_index, = np.where(dist == np.min(dist)) # choose closest prototype
state_index = min_index[0]
return path
def test_distance(self, two_dimensional_points, three_dimensional_points):
assert distance(*two_dimensional_points) > 7.06
assert distance(*two_dimensional_points) < 7.08
def test_distance(self, two_dimensional_points, three_dimensional_points):
assert distance(*two_dimensional_points) > 7.06
assert distance(*two_dimensional_points) < 7.08
def test_distance(self):
sample1 = [0.074936026, 0.0683657238]
center1 = [0.079113924099999997, 0.065822784800000006]
dist = distance(sample1, center1)
self.assertAlmostEqual(dist, 0.00489095)
def rbfMult(x, c, s):
return (1 + (km.distance(x, c))**2)**0.5
def rbfGaus(x, c, s):
return np.exp(-1 / (2 * s**2) * (km.distance(x, c))**2)
def rbfMultIn(x, c, s):
return 1 / (1 + (km.distance(x, c))**2)**0.5