-
Notifications
You must be signed in to change notification settings - Fork 0
/
ham.py
executable file
·267 lines (220 loc) · 10 KB
/
ham.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
# Opting for a procedural approach
from util import *
from news import News
from sklearn import svm
from sklearn import naive_bayes
from sklearn.linear_model import SGDClassifier
from sklearn.ensemble import GradientBoostingClassifier
from sklearn.metrics import *
import matplotlib.pyplot as plt
import math
from random import choice
import numpy as np
from nltk.corpus import stopwords
from sklearn.feature_extraction.text import CountVectorizer, TfidfVectorizer
from nltk.stem.wordnet import WordNetLemmatizer
# from output import *
class HAM(object):
def __init__(self, model, vectorizer):
self.model = model
self.vectorizer = vectorizer
self.news_market_data = False
self.movie_review_data = False
self.models = []
self.pos_acc = []
self.neg_acc = []
self.avg_acc = []
self.pos_prec = []
self.neg_prec = []
self.avg_prec = []
self.pos_rec = []
self.neg_rec = []
self.avg_rec = []
self.pos_f1 = []
self.neg_f1 = []
self.avg_f1 = []
def train_test(self, name="A Model", charts=False):
self.model.fit(self.train_vecs, self.train_labs)
preds = self.model.predict(self.test_vecs)
self.test_labs
if charts is True:
self.graph_data(self.test_labs, preds, name)
else:
print "Test %s", name
print classification_report(self.test_labs, preds,
[-1, 1],
['Negative', 'Positive'])
def graph_data(self, labs, preds, model="A Model"):
pos_acc = accuracy_score(labs, preds)
neg_acc = accuracy_score(labs, preds)
avg_acc = accuracy_score(labs, preds)
pos_prec = precision_score(labs, preds, labels=[-1, 1], pos_label=1)
neg_prec = precision_score(labs, preds, labels=[-1, 1], pos_label=-1)
avg_prec = precision_score(labs, preds, labels=[-1, 1], pos_label=None, average="weighted")
pos_rec = recall_score(labs, preds, labels=[-1, 1], pos_label=1)
neg_rec = recall_score(labs, preds, labels=[-1, 1], pos_label=-1)
avg_rec = recall_score(labs, preds, labels=[-1, 1], pos_label=None, average="weighted")
pos_f1 = f1_score(labs, preds, labels=[-1, 1], pos_label=1)
neg_f1 = f1_score(labs, preds, labels=[-1, 1], pos_label=-1)
avg_f1 = f1_score(labs, preds, labels=[-1, 1], pos_label=None, average="weighted")
print "\n\n"
output = '\\begin{tabular}{c | c c c c}\n'
output += "\\textbf{%s}\t& Accuracy\t& Precision\t& Recall\t& F1 Score\t\\\\\n" % (model)
output += "\\hline \n"
output += "Negative\t\t& %.3f\t\t& %.3f\t\t& %.3f\t\t& %.3f\t\t\\\\\n" % (neg_acc, neg_prec, neg_rec, neg_f1)
output += "Positive\t\t& %.3f\t\t& %.3f\t\t& %.3f\t\t& %.3f\t\t\\\\\n" % (pos_acc, pos_prec, pos_rec, pos_f1)
output += "Average \t\t& %.3f\t\t& %.3f\t\t& %.3f\t\t& %.3f\t\t\\\\\n" % (pos_acc, avg_prec, avg_rec, avg_f1)
output += "\\end{tabular}"
print output
self.models.append(model)
self.pos_acc.append(pos_acc)
self.neg_acc.append(neg_acc)
self.avg_acc.append(avg_acc)
self.pos_prec.append(pos_prec)
self.neg_prec.append(neg_prec)
self.avg_prec.append(avg_prec)
self.pos_rec.append(pos_rec)
self.neg_rec.append(neg_rec)
self.avg_rec.append(avg_rec)
self.pos_f1.append(pos_f1)
self.neg_f1.append(neg_f1)
self.avg_f1.append(avg_f1)
def plot_charts(self):
plot_chart(self.avg_acc, self.models, name="accuracy", title="Accuracy", yaxis="Accuracy (%)")
plot_chart_3(self.pos_prec, self.neg_prec, self.avg_prec, self.models, name="precision", title="Precision", yaxis="Precision (%)")
plot_chart_3(self.pos_rec, self.neg_rec, self.avg_rec, self.models, name="recall", title="Recall", yaxis="Recall (%)")
plot_chart_3(self.pos_f1, self.neg_f1, self.avg_f1, self.models, name="f1", title="F1 Score", yaxis="F1 Score (%)")
def summary_chart(self):
print "\n\n"
output = '\\begin{tabular}{c | c c c c}\n'
output += "\\textbf{%s}\t& Accuracy\t& Precision\t& Recall\t& F1 Score\t\\\\\n" % ("Summary")
output += "\\hline \n"
for i in range(len(self.models)):
output += "%s\t\t& %.3f\t\t& %.3f\t\t& %.3f\t\t& %.3f\t\t\\\\\n" % (self.models[i], self.avg_acc[i], self.avg_prec[i], self.avg_rec[i], self.avg_f1[i])
output += "\\end{tabular}"
print output
def print_doc_feats(self):
for feature in self.vectorizer.get_feature_names():
print feature
def prep_news_data(self):
if not self.news_market_data:
print 'Preparing news and stock data...\n'
news = News('Resources/articles.db')
raw = news.db_articles()
train_raw, test_raw = divide_list_by_ratio(raw) # prep_news_data returns a tuple of vectors, labels
self.train_vecs, self.train_labs = self.prep_news_articles(train_raw, fit=True)
self.test_vecs, self.test_labs = self.prep_news_articles(test_raw)
self.news_market_data = True
self.movie_review_data = False
def prep_news_articles(self, raw_docs, fit=False):
docs = self.filter_old_news(raw_docs)
doc_labels = self.news_labels(docs)
doc_bodies = map(lambda x: x[4], docs) # 3 is title, 4 is body
if fit:
self.vectorizer.fit(doc_bodies, doc_labels)
doc_vectors = self.vectorizer.transform(doc_bodies)
if isinstance(self.model, naive_bayes.GaussianNB): # check if need dense vectors
doc_vectors = doc_vectors.toarray()
return doc_vectors, doc_labels
def prep_reviews_data(self): # messy code to test classifier with movie reviews
if not self.movie_review_data:
print 'Preparing movie reviews...\n'
from nltk.corpus import movie_reviews
docs = [movie_reviews.raw(fileid)
for category in movie_reviews.categories()
for fileid in movie_reviews.fileids(category)]
process = lambda x: 1 if x == 'pos' else -1
labels = [process(category)
for category in movie_reviews.categories()
for fileid in movie_reviews.fileids(category)]
docs, labels = double_shuffle(docs, labels)
training, testing = divide_list_by_ratio(docs)
self.train_labs, self.test_labs = divide_list_by_ratio(labels)
train_vecs = self.vectorizer.fit_transform(training)
test_vecs = self.vectorizer.transform(testing)
if isinstance(self.model, naive_bayes.GaussianNB):
train_vecs = train_vecs.toarray()
test_vecs = test_vecs.toarray()
self.train_vecs = train_vecs
self.test_vecs = test_vecs
self.movie_review_data = True
self.news_market_data = False
def filter_old_news(self, docs):
fn = lambda d: bool(db_symbol_change(d[2], datetime.strptime(d[0], '%Y-%m-%d')))
return filter(fn, docs)
def news_labels(self, corpus):
''' Returns a numpy array of integer labels that correspond to the corpus docs.
1 for a doc about a stock that happened to go up, -1 for a doc about a stock that went down. Removes data entry if no stock data.
'''
labels = []
for doc in corpus:
date = datetime.strptime(doc[0], '%Y-%m-%d')
change = db_symbol_change(doc[2], date)
if change:
label = change/math.fabs(change) # => 1 or -1
labels.append(label)
return np.array(labels, dtype=np.int8)
def advanced_train_test():
self.model.fit(self.train_vecs, self.train_labs)
preds = model.predict(self.test_vecs)
total = len(preds)
correct = 0.0
up_count = 0
down_count = 0
for pred, act in zip(preds, self.test_labs):
if pred == 1:
up_count += 1
elif pred == -1:
down_count += 1
if pred == act:
correct += 1
acc = correct/total
print "%d/%d Correct" % (correct, total)
print "Accuracy: %.2f" % acc
print "Predictions: %d UP, %d DOWN" % (up_count, down_count)
return acc
>>>>>>> 7570875e948014a1e5921d292902569b750e17ee
class RandomClassifier(object):
def fit(self, A, B):
pass
def predict(self, A):
return [choice([-1, 1]) for i in range(len(A))]
# class DocPreprocessor(object):
# def __init__(self):
# self.wnl = WordNetLemmatizer()
# def preprocess(self, doc):
# def process_word(word):
# word = word.lower()
# word = self.wnl.lemmatize(word)
# return word
# return ' '.join(map(process_word, doc.split()))
def plot_chart(results, labels, name="figure", title=None, yaxis=None):
N = len(results)
ind = np.arange(N)
width = 0.35
fig = plt.figure(facecolor="#ffffff")
ax = fig.add_subplot(111)
rects1 = ax.bar(ind + (width / 4), results, width, color='#6AA8EB')
# add some
ax.set_ylabel(yaxis, family="serif")
ax.set_title(title, family="serif")
ax.set_xticks(ind + (width * 3 / 4))
ax.set_xticklabels(labels, family="serif", fontname="Computer Modern", size="x-small", rotation=14)
plt.savefig(name + '.png')
# plt.show()
def plot_chart_3(results, results2, results3, labels, name="figure", title=None, yaxis=None):
N = len(results)
ind = np.arange(N)
width = 0.30
fig = plt.figure(facecolor="#ffffff")
ax = fig.add_subplot(111)
pos = ax.bar(ind, results, width, color='#C7430A')
neg = ax.bar(ind + width, results2, width, color='#456772')
avg = ax.bar(ind + 2 * width, results3, width, color='#378F00')
# add some
ax.set_ylabel(yaxis, family="serif")
ax.set_title(title, family="serif")
ax.set_xticks(ind + 3 * width / 2)
ax.set_xticklabels(labels, family="serif", fontname="Computer Modern", size="x-small", rotation=14)
ax.legend((pos, neg, avg), ('Positive', 'Negative', 'Average'))
plt.savefig(name + '.png')