-
Notifications
You must be signed in to change notification settings - Fork 0
/
read_tree.py
278 lines (238 loc) · 9.45 KB
/
read_tree.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
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
from sklearn import datasets
from sklearn.cluster import AgglomerativeClustering
from numpy import *
import time
from numpy import *
import sys
sys.setrecursionlimit(12000)
def get_news_group_pca(dimension):
from sklearn.datasets import fetch_20newsgroups
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.decomposition import PCA
newsgroups_train = fetch_20newsgroups(subset='train')
vectorizer = TfidfVectorizer(min_df=0.01, max_df=0.95)
train_data = vectorizer.fit_transform(newsgroups_train.data)
train_data = train_data.todense()
train_labels = newsgroups_train.target;
# initialise PCA with n_components = dimension
pca = PCA(n_components=dimension)
# apply pca
pca.fit(train_data)
Y = pca.transform(train_data)
n_elements = len(unique(train_labels))
return Y, n_elements, train_labels, len(set(train_labels))
def get_news_group(size):
from sklearn.datasets import fetch_20newsgroups
from sklearn.feature_extraction.text import TfidfVectorizer
newsgroups_train = fetch_20newsgroups(subset='train')
vectorizer = TfidfVectorizer(min_df=0.01, max_df=0.95)
train_data = vectorizer.fit_transform(newsgroups_train.data)
train_data = train_data.todense()
train_labels = newsgroups_train.target;
# initialise PCA with n_components = dimension
n_elements = len(unique(train_labels[0:size]))
return train_data[0:size, :], n_elements, train_labels[0:size], len(set(train_labels[0:size]))
def read_file(filename):
f = open(filename, "r")
lines = f.readlines()
n = len(lines)+1
nb_clust = n
#print(n)
clusters = {3*i*n+1: i for i in range(n)}
T = [[i,-1] for i in range(n)]
for l in lines:
words = l.split(";")
res, c1, c2 = words[:3]
idres_str = res.split(",")
idres = int(idres_str[0])*3*n + int(idres_str[1])
idc1_str = c1.split(",")
idc1 = int(idc1_str[0])*3*n + int(idc1_str[1])
idc2_str = c2.split(",")
idc2 = int(idc2_str[0])*3*n + int(idc2_str[1])
#print(idres_str, idc2_str, idc1_str)
#print(idres, idc1, idc2)
clusters[idres] = nb_clust
T.append([clusters[idc1], clusters[idc2]])
nb_clust+=1
return T
def get_cluster(tree, index):
if tree[index][0] == index:
return [index]
return get_cluster(tree, tree[index][0]) + get_cluster(tree, tree[index][1])
def clusters(tree, k):
nums = []
i = 1
while len(nums) != k:
current = len(tree)-i
if current in nums:
nums.remove(current)
c1 = tree[-i][0]
c2 = tree[-i][1]
nums.append(c1)
if c1 != current:
nums.append(c2)
i+=1
return [get_cluster(tree, nums[i]) for i in range(k)]
def get_dataset(name):
from sklearn.preprocessing import scale
data = []
if name == "cancer":
from sklearn.datasets import load_breast_cancer
dataset = load_breast_cancer()
elif name == "digits":
from sklearn.datasets import load_digits
dataset = load_digits()
elif name == "iris":
from sklearn.datasets import load_iris
dataset = load_iris()
elif name == "boston":
from sklearn.datasets import load_boston
dataset = load_boston()
elif name == "KDD":
from sklearn.datasets import fetch_kddcup99
dataset = fetch_kddcup99(subset='SF')
data = dataset.data[:2000, [0,2,3]]
else:
print("Unknown name of dataset")
exit(-1)
labels = dataset.target
if data == []:
data = scale(dataset.data)
n_samples, n_features = data.shape
n_elements = len(unique(labels))
return data, n_elements, labels, len(set(labels))
# example
# t = [[0, -2], [1, -1], [2, -1], [3, -1], [4, -1], [0, 1], [2, 3], [5, 6], [7, 4]]
# print(clusters(t, 3))
#from sklearn.datasets import load_iris
from sklearn.metrics.cluster import normalized_mutual_info_score
def convert(clusters, n):
clustering_vect = [0]*n
for i in range(len(clusters)):
for p in clusters[i]:
clustering_vect[p] = i
return clustering_vect
def approx_vs_ward(e, number_of_visited_leafs, numebr_of_trees, name, dimension, size):
#print("epsilon ", "0." + str(epsilon), numebr_of_trees, number_of_visited_leafs)
# #eps = [25, 50, 75, 85, 95, 200, 400];
output_file = 'result_epsilons.txt'
with open(output_file, 'a') as file:
file.write('epsilon 0.' + str(e) + ' ' + str(numebr_of_trees) + ' ' + str(number_of_visited_leafs) + ' ' + str(size) + ' ' + str(dimension) +'\n')
data, n, labels, k = get_news_group(size)
file_name = './' + name + '_' + str(dimension) + '_' + str(e) + '_' + str(numebr_of_trees) + '_' + str(number_of_visited_leafs) + ".out"
T = read_file(file_name)
print (k)
clust = clusters(T, k)
print (len(clust))
file.write('Algo ' + str(normalized_mutual_info_score(convert(clust, len(labels)), labels)) + '\n')
ward = AgglomerativeClustering(n_clusters=k, linkage='ward', connectivity=None)
data = loadtxt('news_' + str(size) + '.in', skiprows = 1)
print (data.shape)
clustering = ward.fit(data)
clust = clustering.labels_
file.write('std_ward ' + str(normalized_mutual_info_score(clust, labels)) + '\n')
#
# for name in data_sets:
# data, n, labels, k = get_dataset(name)
# ward = AgglomerativeClustering(n_clusters=k, linkage='ward', connectivity=None)
# clustering = ward.fit(data)
# clust = clustering.labels_
# file.write('Ward ' + str(normalized_mutual_info_score(clust, labels)) + '\n')
#
# for name in data_sets:
# data, n, labels, k = get_dataset(name)
# ward = AgglomerativeClustering(n_clusters=k, linkage='average', connectivity=None)
# clustering = ward.fit(data)
# clust = clustering.labels_
# file.write('Average ' + str(normalized_mutual_info_score(clust, labels)) + '\n')
#
# for name in data_sets:
# data, n, labels, k = get_dataset(name)
# ward = AgglomerativeClustering(n_clusters=k, linkage='single', connectivity=None)
# clustering = ward.fit(data)
# clust = clustering.labels_
# file.write('Single ' + str(normalized_mutual_info_score(clust, labels)) + '\n')
def readFILE(file_name):
with open(file_name) as f:
content = f.readline()
content = content.split(' ')
n = int(content[0])
d = int(content[1])
#k = int(content[2])
X = zeros((n, d))
for i in range(n):
content = f.readline()
content = content.split(' ')
for j in range(d):
X[i, j] = float(content[j])
return X
#
# std::vector<int> trees = {4 , 16};
# std::vector<int> leaves = {5, 128};
# std::vector<float> epsilons = {0.5, 1, 7};
#trees = [2]
#leaves = [10]
#epsilons = [800]
#d = 20
#k = 10
#ward = AgglomerativeClustering(n_clusters=k, linkage='ward', connectivity=None)
# with open('ward_accuracy7.txt', 'w') as f:
# for e in epsilons:
# for t in trees:
# for l in leaves:
# for i in range(10000, 20000, 1000):
# for j in range(1):
# if i == 18000:
# j = 1
# data_file = './data/data' + str(i) + '_' + str(j) + '_' + str(d) + '_' + str(k) + '.in'
# data = readFILE(data_file)
# start = time.time()
# clustering = ward.fit(data)
# end = time.time()
# labels = clustering.labels_
# res_file = 'data' + str(i) + '_' + str(j) + '_' + str(d) + '_' + str(k) + '_' + str(e) + '_' + str(t) + '_' + str(l) + '.out'
# T = read_file(res_file)
# clust = clusters(T, k);
# acc = normalized_mutual_info_score(convert(clust, len(labels)), labels)
# print(str(end - start))
# f.write(str(end - start) + ' ' + str(acc) + ' ')
# f.write('\n')
#
#data = readFILE("./data/iris.in")
#clustering = ward.fit(data)
#random.shuffle(data)
#labels = clustering.labels_
# data, n, labels, k = get_dataset('boston')
# file_name = './data/boston800_2_10.out'
# T = read_file(file_name)
# clust = clusters(T, k)
# print(normalized_mutual_info_score(convert(clust, len(labels)), labels))
#data, n, labels, k = get_dataset('boston')
#print(k)
## e psilon, number_of_visited_leafs, number_of_trees, dimension
def exp_sizes_acc():
sizes = 100
while sizes < 11314:
approx_vs_ward(100, 128, 25, 'news' + str(sizes), 2164, sizes)
sizes = sizes * 2
def tree_sizes_perf():
sizes = 11314
tree = [1, 4, 8, 12, 16, 20, 24, 28, 32]
for t in tree:
approx_vs_ward(200, 128, t, 'news' + str(sizes), 2164, sizes)
def leaves_perf():
sizes = 11314
leaves = [2, 4, 8, 16, 32, 64, 128, 256, 512, 1024]
for l in leaves:
approx_vs_ward(200, l, 16, 'news' + str(sizes), 2164, sizes)
def epsilons_perf():
sizes = 11314
epsilons = [50, 100, 200, 400, 800, 1000]
for e in epsilons:
approx_vs_ward(e, 32, 16, 'news' + str(sizes), 2164, sizes)
epsilons_perf()
#leaves_perf()
#tree_sizes_perf()
#_, _, labels, _ = get_news_group(2)
#for u in labels:
# print (u)