-
Notifications
You must be signed in to change notification settings - Fork 0
/
grid_search.py
285 lines (249 loc) · 8.11 KB
/
grid_search.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
279
280
281
282
283
284
285
from numpy.lib.shape_base import vstack, hstack
from optparse import OptionParser
from result_util import get_fscore_output
import datetime
import pickle
from common import FILES_PATH, RESULT_FILES_PATH
import StringIO
import sys
print __doc__
from pprint import pprint
import numpy as np
from sklearn import datasets
from sklearn import preprocessing
from sklearn.cross_validation import StratifiedKFold
from sklearn.grid_search import GridSearchCV
from sklearn.metrics import classification_report
from sklearn.metrics import precision_score
from sklearn.metrics import recall_score
from sklearn.metrics import f1_score
from sklearn.svm import SVC
from sklearn.datasets import load_svmlight_file
from numpy.core.numeric import array
class ob:
def __init__(self):
self.stream = StringIO.StringIO()
def write(self, data):
if not isinstance(data, unicode):
data = data.decode('utf-8')
self.stream.write(data.encode('utf-8'))
self.stream.flush()
def __getattr(self, attr):
return getattr(self.stream, attr)
def getvalue(self):
return self.stream.getvalue()
def flush(self):
self.write(self.getvalue())
def load_data(filename, nparray=True, feature_list=None):
"""
Returns (features [m x n], labels [m x 1])
filename - filename to parse
nparray - return numpy array
feature_list - a list of feature numbers (which are used in the file) to load. If None, all the features are loaded
"""
X = []
Y = []
i = 0
f = open(filename, 'r')
if len(feature_list) == 0: feature_list = None
for l in f:
i += 1
tokens = l.split(' ')
curx = []
if len(tokens) > 1:
Y.append(int(tokens[0]))
for t in tokens[1:]:
tokens2 = t.split(':')
if len(tokens2) == 2:
if len(feature_list) and float(tokens2[0]) not in feature_list:
# we skip this feature
continue
x = float(tokens2[1])
else:
x = 0.0
curx.append(x)
X.append(curx)
f.close()
if nparray:
return (array(X), array(Y))
return (X, Y)
global_time = datetime.datetime.today()
def log(msg):
diff = datetime.datetime.today() - global_time
print datetime.datetime.today().strftime('%Y-%m-%d %H:%M:%S'), diff, msg
################################################################################
# Loading the Digits dataset
#digits = datasets.load_digits()
# To apply an classifier on this data, we need to flatten the image, to
# turn the data in a (samples, feature) matrix:
#n_samples = len(digits.images)
#X = digits.images.reshape((n_samples, -1))
#y = digits.target
op = OptionParser()
op.add_option('-p',
'--prefix',
help='training and test file prefix (without .train/.test)',
dest='prefix',
default=None)
op.add_option('-d',
'--data',
help='training and test file (will be divided into training/validation/test sets',
dest='data',
default=None)
op.add_option('-r',
'--results',
help='results file. if omitted, data filename + date will be used',
dest='results',
default=None)
op.add_option('-f',
'--features',
help='which features to use (use the number in the file)',
dest='features',
default=None)
(options, args) = op.parse_args()
prefix = 'tallinn_200s_title_rdist_1000'
prefix = 'data/tallinn_111_titlewc_rdist_10000'
prefix = 'data/tallinn_202s_edistsim_rdist_10000'
prefix = 'data/tallinn_200ss_title_rdist_1000'
bln_write_model = True
testX = None
feature_list = []
if options.features:
print options.features
feature_list = map(lambda s : int(s), options.features.split(','))
print feature_list
if options.prefix:
prefix = options.prefix
#prefix = 'data/tallinn_201s_title_rdist_10000'
X, y = load_data(prefix + '.train', feature_list=feature_list)
#print y.shape
#X, y = load_svmlight_file(prefix + '.train')
testX, testY = load_data(prefix + '.test', feature_list=feature_list)
X = vstack((X, testX))
X = preprocessing.scale(X)
print X
#y = vstack((y, testY))
#y = y + testY
#y = y.transpose()
#print y.shape, testY.shape
y = hstack((y, testY))
print y
if options.data:
log('loading ' + str(options.data))
X, y = load_data(options.data, feature_list=feature_list)
if False:
pos_start = 0
pos_end = 1500
neg_start = 2000
neg_end = 1000000
# tmp
tmp = X[pos_start:pos_end, :]
tmp = vstack((tmp, X[neg_start:neg_end]))
X = tmp
tmp = y[pos_start:pos_end]
tmp = hstack((tmp, y[neg_start:neg_end]))
y = tmp
#X = X[0:2000, :]
#y = y[0:2000]
#print X
#print y
#exit()
#pprint(np.array(X))
#pprint(y)
# split the dataset in two equal part respecting label proportions
train, test = iter(StratifiedKFold(y, 2)).next()
print X[train].shape, y[train].shape
log('start training')
################################################################################
# Set the parameters by cross-validation
tuned_parameters = [{'kernel': ['rbf'], 'gamma': [1e-3, 1e-4],
'C': [1, 10, 100, 1000]},
{'kernel': ['linear'], 'C': [1, 10, 100, 1000]}]
scores = [
#('precision', precision_score),
#('recall', recall_score),
('f1_score', f1_score),
]
for score_name, score_func in scores:
clf = GridSearchCV(SVC(C=1, cache_size=100), tuned_parameters, score_func=score_func, verbose=2, n_jobs=8)
clf.fit(X[train], y[train], cv=StratifiedKFold(y[train], 5))
y_true, y_pred = y[test], clf.predict(X[test])
sys.stdout = ob()
log('score_name '+ str(score_name))
print "Classification report for the best estimator: "
print clf.best_estimator
print "Tuned for '%s' with optimal value: %0.3f" % (
score_name, score_func(y_true, y_pred))
print classification_report(y_true, y_pred)
print "Grid scores:"
pprint(clf.grid_scores_)
print
if testX is not None:
# let's use clf:
predict_result = clf.predict(testX)
i = 0
tp = 0
fp = 0
fn = 0
for p in predict_result:
if p == testY[i]:
if p == 1:
tp += 1
else:
# p != Y
if p == 1:
fp += 1
else:
fn += 1
i += 1
s = get_fscore_output(tp, fp, fn)
print s
print
"""
Let's test the whole set (training + test)
"""
predict_result = clf.predict(X)
i = 0
tp = 0
fp = 0
fn = 0
for p in predict_result:
if p == y[i]:
if p == 1:
tp += 1
else:
# p != Y
if p == 1:
fp += 1
else:
fn += 1
i += 1
s = get_fscore_output(tp, fp, fn)
print s
str_output = sys.stdout.getvalue()
sys.stdout = sys.__stdout__
print str_output
if bln_write_model:
strtime = datetime.datetime.today().strftime('%Y-%m-%d-%H-%M-%S')
mname = 'x'
if options.prefix:
mname = options.prefix
elif options.data:
dotpos = options.data.find('.')
if dotpos > 0:
mname = options.data[:dotpos]
slashpos = mname.rfind('/')
if slashpos > 0:
mname = mname[slashpos+1:]
fname = 'model_' + mname + '_' + str(strtime) + '.pickle'
pickle.dump(clf.best_estimator, open(FILES_PATH + '/' + fname, 'w'))
# results
if options.results:
fname = options.results
else:
fname = RESULT_FILES_PATH + '/' + 'results_' + mname + '_' + str(strtime) + '.results'
fresults = open(fname, 'w')
fresults.write(str_output)
fresults.close()
#return (tp, fp, fn, predict_result, clf)
log('done')