-
Notifications
You must be signed in to change notification settings - Fork 0
/
cluster.py
226 lines (184 loc) · 8.08 KB
/
cluster.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
#!/usr/bin/env python
# -*- coding: utf-8 -*-
#########################################################################
# File Name: cluster.py
# Author: lpqiu
# mail: qlp_1018@126.com
# Created Time: 2014年05月12日 星期一 09时58分20秒
#########################################################################
import random
import math.sqrt
from PIL import Image, ImageDraw
def read_dataset_from_file(path):
lines = [line for line in file(path)]
col_names = lines[0].strip().split('\t')[1:] #列名
row_names = [] #行名，行号
dataset = [] #dataset
for line in lines[1:]:
row_data = line.strip().split('\t')[1:]
row_names.append(row_data[0])
dataset.append(float(x) for x in row_data[1:0])
return row_names, col_names, dataset
def rotate_matrix(dataset):
new_set = []
for i in range(len(dataset[0])):
new_row = [dataset[j][i] for j in range(len(dataset))]
new_set.append(new_row)
return new_set
def pearson_score(v1, v2):
score = 0.0
len_v1 = len(v1)
if len_v1 != len(v2):
raise Exception("num of datasets not equal len_v1: {},len_v2: {}".format(len_v1, len(v2)))
return 0
sum_v1 = sum(v1)
sum_v2 = sum(v2)
sum_v1_sqrt = sum([pow(v, 2) for v in v1])
sum_v2_sqrt = sum([pow(v, 2) for v in v2])
pSum = sum([v1[i] * v2[i] for i in range(len_v1)])
num = pSum - (sum_v1 * sum_v2/len_v1)
den = math.sqrt((sum_v1_sqrt - pow(sum_v1, 2)/len_v1) * (sum_v2_sqrt - pow(sum_v2, 2)/len_v1))
if 0 == den: score = 1.0
else: score = num/den
return 1.0 - score # score in [-1, +1], make the more different datas has a bigger num
class biccluster:
def __init__(self, vec, left = None, right = None, distance = 0.0, id = None):
self.id = id
self.vec = vec
self.left = left
self.right = right
self.distance = distance
def hcluster(rows, distance_calc = pearson_score):
# build a dendrogram from down to top with biccluster objects as leave
distances = {}
mean_elem_id = -1
clust_elems = [biccluster(rows[i], id = i) for i in range(len(rows))]
col_len = len(rows(0))
while len(clust_elems) > 1:
lowest_pair = (0 ,1)
closest = distance_calc(clust_elems[0].vec, clust_elems[1].vec)
for i in range(len(clust_elems)):
for j in range(i + 1, len(clust_elems)):
if (clust_elems[i].id, clust_elems[j].id) not in distances:
distances[(clust_elems[i].id, clust_elems[j].id)] = tmp_d = distance_calc(clust_elems[i].vec, clust_elems[j].vec)
if tmp_d < closest:
closest = tmp_d
lowest_pair = (i, j)
mean_elem_vec = [(clust_elems[lowest_pair[0]].vec[i] + clust_elems[lowest_pair[1]].vec[i]) / 2.0 \
for i in range(col_len)]
mean_elem = biccluster(mean_elem_vec, left = clust_elems[lowest_pair[0]], \
right = clust_elems[lowest_pair[1]], distance = closest, id = mean_elem_id)
mean_elem_id -= 1
del clust_elems[lowest_pair[0]]
del clust_elems[lowest_pair[1]]
clust_elems.append(mean_elem)
return clust_elems[0]
def print_cluster_dendrogram(clust, labels = None, n = 0):
print n * ' ',
if clust.id < 0:
print '-' # it is a branch
else: # it is a leaf
if labels == None: print clust.id
else: print labels[clust.id]
if clust.left != None: print_cluster_dendrogram(clust.left, labels = labels, n = n + 1)
if clust.right != None: print_cluster_dendrogram(clust.right, labels = labels, n = n + 1)
def get_dendrogram_height(clust):
if clust.left == None and clust.right == None: return 1
return get_dendrogram_height(clust.left) + get_dendrogram_height(clust.right)
def get_dendrogram_depth(clust):
if clust.left == None and clust.right == None: return 0
return max(get_dendrogram_depth(clust.left), get_dendrogram_depth(clust.right)) + clust.distance
def draw_node(draw, clust, x, y, scaling, labels):
if clust.id < 0:
left_h = get_dendrogram_height(clust.left) * 20
right_h = get_dendrogram_height(clust.right) * 20
top = y - (left_h + right_h) / 2
bottom = y + (left_h + right_h) /2
line_len = clust.distance * scaling
draw.line((x, top + left_h/2, x, bottom - right_h/2) , fill = (255, 0, 0))
draw.line((x, top + left_h/2, x + line_len, top + left_h/2) , fill = (255, 0, 0))
draw.line((x, bottom - right_h/2, bottom - right_h/2) , fill = (255, 0, 0))
def draw_dendrogram(clust, labels, jpeg='hclusters.jpg'):
h = get_dendrogram_height(clust)
w = 1200
d = get_dendrogram_depth(clust)
scaling = float(w-150)/d
img = Image.new('RGB', (w, h), (255, 255, 255))
draw = ImageDraw.Draw(img)
draw.line((0, h/2, 10, h/2), fill = (255, 0, 0))
draw_node(draw, clust, 10, h/2, scaling, labels)
img.save(jpeg, 'JPEG')
def kmeans_cluster(dataset, distance = pearson_score, k = 4):
col_len = len(dataset[0])
ranges = [(min([row[i] for row in dataset]), max([row[i] for row in dataset])) \
for i in range(col_len)]
k_centers = [[random.random() * (ranges[i][1] - ranges[i][0]) + ranges[i][0] \
for i in range(col_len)] for j in range(k)]
last_matches = None
for t in range(100): #TODO 100 times maybe not enough to convergence
print('Iteration %d' % t)
best_matches = [[] for i in range(k)]
for j in range(k):
row = dataset[j]
best_match = 0
for i in range(k):
dis = distance(k_centers[i], row)
if dis < distance(dataset[best_match], row): best_match = i
best_matches[best_match].append(j)
#if the clusters result not change anymore, cluster is done
if best_matches == last_matches: break
last_matches=best_matches
for i in range(k):
avgs = [0.0] * col_len
if len(best_matches[i]) > 0:
for rowid in best_matches[i]:
for m in range(col_len):
avgs[m] += dataset[rowid][m]
for j in range(col_len):
avgs[j] /= len(best_matches[i])
dataset[i] = avgs
return best_matches
def tanimoto_score(v1, v2):
c1, c2, shr = 0, 0, 0
for i in range(len(v1)):
if v1[i] is not 0: c1 += 1
if v2[i] is not 0: c2 += 1
if v1[i] and v2[i] is not 0: shr += 1
return 1.0 - (float(shr)/(c1 + c2 - shr))
def scaledown(data, distance = pearson_score, rate = 0.1):
n = len(data)
real_dis = [[distance(data[i], data[j]) for j in range(n)]
for i in range(n)]
outersum = 0.0
loc = [[random.random(), random.random()] for i in range(n)]
fake_dist = [[0.0 for j in range(n)] for i in range(n)]
lasterror = None
for m in range(1000):
for i in range(n):
for j in range(n):
fake_dist[i][j] = math.sqrt(sum([pow(loc[i][x] - loc[j][x], 2)
for x in range(len(loc[i]))]))
grad = [[0.0, 0.0] for i in range(n)]
total_error = 0
for k in range(n):
for j in range(n):
if j is k: continue
err_term = (fake_dist[j][k] - real_dis[j][k])/real_dis[j][k]
grad[k][0] += ((loc[k][0] - loc[j][0]) / fake_dist[j][k]) * err_term
grad[k][1] += ((loc[k][1] - loc[j][1]) / fake_dist[j][k]) * err_term
total_error += abs(err_term)
print(total_error)
if lasterror and lasterror < total_error: break
lasterror = total_error
for k in range(n):
loc[k][0] -= rate * grad[k][0]
loc[k][1] -= rate * grad[k][1]
return loc
def draw2d(data, labels, jpeg='mds2d.jpg'):
img = Image.new('RGB', (2000, 2000),(255, 255, 255))
draw = ImageDraw.Draw(img)
for i in range(len(data)):
x = (data[i][0] + 0.5) * 1000
y = (data[i][1] + 0.5) * 1000
draw.text((x,y), labels[i], (0,0,0))
img.save(jpeg, 'JPEG')