-
Notifications
You must be signed in to change notification settings - Fork 1
/
advcl_EnsClass_Bagging.py
273 lines (234 loc) · 7.25 KB
/
advcl_EnsClass_Bagging.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
import matplotlib.pyplot as plt
from sklearn.ensemble import BaggingClassifier, RandomForestClassifier
from sklearn.metrics import (
accuracy_score,
auc,
classification_report,
f1_score,
plot_confusion_matrix,
roc_auc_score,
roc_curve,
)
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import LabelBinarizer, LabelEncoder
import utils
def draw_confusion_matrix(Clf, X, y):
titles_options = [
("Confusion matrix, without normalization", None),
("Bagging Confusion matrix", "true"),
]
for title, normalize in titles_options:
disp = plot_confusion_matrix(Clf, X, y, cmap="Greens", normalize=normalize)
disp.ax_.set_title(title)
plt.show()
"""
# DATASET COMPLETO
df = utils.load_tracks(
"data/tracks.csv", dummies=True, buckets="discrete", fill=True, outliers=True
)
print(df["album", "type"].unique())
# feature to reshape
label_encoders = dict()
column2encode = [
("track", "language_code"),
("album", "listens"),
("album", "type"),
("track", "license"),
("album", "comments"),
("album", "date_created"),
("album", "favorites"),
("artist", "comments"),
("artist", "date_created"),
("artist", "favorites"),
("track", "comments"),
("track", "date_created"),
("track", "duration"),
("track", "favorites"),
("track", "interest"),
("track", "listens"),
]
for col in column2encode:
le = LabelEncoder()
df[col] = le.fit_transform(df[col])
label_encoders[col] = le
df.info()
"""
# DATASET PICCOLINO
df = utils.load_small_tracks(buckets="discrete")
label_encoders = dict()
column2encode = [
("track", "duration"),
("track", "interest"),
("track", "listens"),
("album", "type"),
]
for col in column2encode:
le = LabelEncoder()
df[col] = le.fit_transform(df[col])
label_encoders[col] = le
df.info()
class_name = ("album", "type")
attributes = [col for col in df.columns if col != class_name]
X = df[attributes].values
y = df[class_name]
X_train, X_test, y_train, y_test = train_test_split(
X, y, test_size=0.2, random_state=100, stratify=y
)
"""BAGGING DECISION TREE
clf = BaggingClassifier(
base_estimator=DecisionTreeClassifier(
criterion="gini",
max_depth=9,
min_samples_split=10,
min_samples_leaf=10,
),
n_estimators=100,
random_state=0,
)
clf.fit(X_train, y_train)
# Apply on the training set
print("Apply on the training set: \n")
Y_pred = clf.predict(X_train)
print("Accuracy %s" % accuracy_score(y_train, Y_pred))
print("F1-score %s" % f1_score(y_train, Y_pred, average=None))
print(classification_report(y_train, Y_pred))
# Apply on the test set and evaluate the performance
print("Apply on the test set and evaluate the performance: \n")
y_pred = clf.predict(X_test)
print("Accuracy %s" % accuracy_score(y_test, y_pred))
print("F1-score %s" % f1_score(y_test, y_pred, average=None))
print(classification_report(y_test, y_pred))
draw_confusion_matrix(clf, X_test, y_test)
ROC CURVE
lb = LabelBinarizer()
lb.fit(y_test)
lb.classes_.tolist()
fpr = dict()
tpr = dict()
roc_auc = dict()
by_test = lb.transform(y_test)
by_pred = lb.transform(y_pred)
for i in range(4):
fpr[i], tpr[i], _ = roc_curve(by_test[:, i], by_pred[:, i])
roc_auc[i] = auc(fpr[i], tpr[i])
roc_auc = roc_auc_score(by_test, by_pred, average=None)
plt.figure(figsize=(8, 5))
for i in range(4):
plt.plot(
fpr[i],
tpr[i],
label="%s ROC curve (area = %0.2f)" % (lb.classes_.tolist()[i], roc_auc[i]),
)
plt.plot([0, 1], [0, 1], "k--")
plt.xlim([0.0, 1.0])
plt.ylim([0.0, 1.05])
plt.title("Bagging Decision Tree Roc-Curve")
plt.xlabel("False Positive Rate", fontsize=10)
plt.ylabel("True Positive Rate", fontsize=10)
plt.tick_params(axis="both", which="major", labelsize=12)
plt.legend(loc="lower right", fontsize=7, frameon=False)
plt.show()
BAGGING SVC
# ricorda che ci mette all'incirca mezz 'ora e i risultati non son top perchè
# non riesce a predire bene la classe 1 anche se ha un accuracy alta
# STANDARDIZZO
scaler = StandardScaler()
X_train = scaler.fit_transform(X_train)
X_test = scaler.fit_transform(X_test)
clf = BaggingClassifier(base_estimator=SVC(C=0.1), n_estimators=10, random_state=0)
clf.fit(X_train, y_train)
y_pred = clf.predict(X_test)
print("Accuracy %s" % accuracy_score(y_test, y_pred))
print("F1-score %s" % f1_score(y_test, y_pred, average=None))
print(classification_report(y_test, y_pred))
draw_confusion_matrix(clf, X_test, y_test)
# ROC CURVE
lb = LabelBinarizer()
lb.fit(y_test)
lb.classes_.tolist()
fpr = dict()
tpr = dict()
roc_auc = dict()
by_test = lb.transform(y_test)
by_pred = lb.transform(y_pred)
for i in range(4):
fpr[i], tpr[i], _ = roc_curve(by_test[:, i], by_pred[:, i])
roc_auc[i] = auc(fpr[i], tpr[i])
roc_auc = roc_auc_score(by_test, by_pred, average=None)
plt.figure(figsize=(8, 5))
for i in range(4):
plt.plot(
fpr[i],
tpr[i],
label="%s ROC curve (area = %0.2f)" % (lb.classes_.tolist()[i], roc_auc[i]),
)
plt.plot([0, 1], [0, 1], "k--")
plt.xlim([0.0, 1.0])
plt.ylim([0.0, 1.05])
plt.title("BAGGING SVM Roc-Curve")
plt.xlabel("False Positive Rate", fontsize=10)
plt.ylabel("True Positive Rate", fontsize=10)
plt.tick_params(axis="both", which="major", labelsize=12)
plt.legend(loc="lower right", fontsize=7, frameon=False)
plt.show()
"""
"""BAGGING RANDOM FOREST"""
# è un pò meglio ma ci mette pure lui mezz'ora ad eseguire
# non è migliore del Boosting RF
clf = BaggingClassifier(
base_estimator=RandomForestClassifier(
n_estimators=100,
criterion="gini",
max_depth=17,
min_samples_split=3,
min_samples_leaf=3,
max_features="auto",
random_state=10,
class_weight="balanced",
),
n_estimators=100,
random_state=0,
)
clf.fit(X_train, y_train)
# Apply on the training set
print("Apply on the training set: \n")
Y_pred = clf.predict(X_train)
print("Accuracy %s" % accuracy_score(y_train, Y_pred))
print("F1-score %s" % f1_score(y_train, Y_pred, average=None))
print(classification_report(y_train, Y_pred))
# Apply on the test set and evaluate the performance
print("Apply on the test set and evaluate the performance: \n")
y_pred = clf.predict(X_test)
print("Accuracy %s" % accuracy_score(y_test, y_pred))
print("F1-score %s" % f1_score(y_test, y_pred, average=None))
print(classification_report(y_test, y_pred))
draw_confusion_matrix(clf, X_test, y_test)
# ROC CURVE
lb = LabelBinarizer()
lb.fit(y_test)
lb.classes_.tolist()
fpr = dict()
tpr = dict()
roc_auc = dict()
by_test = lb.transform(y_test)
by_pred = lb.transform(y_pred)
for i in range(4):
fpr[i], tpr[i], _ = roc_curve(by_test[:, i], by_pred[:, i])
roc_auc[i] = auc(fpr[i], tpr[i])
roc_auc = roc_auc_score(by_test, by_pred, average=None)
plt.figure(figsize=(8, 5))
for i in range(4):
plt.plot(
fpr[i],
tpr[i],
label="%s ROC curve (area = %0.2f)" % (lb.classes_.tolist()[i], roc_auc[i]),
)
plt.plot([0, 1], [0, 1], "k--")
plt.xlim([0.0, 1.0])
plt.ylim([0.0, 1.05])
plt.title("Bagging Random Forest Roc-Curve")
plt.xlabel("False Positive Rate", fontsize=10)
plt.ylabel("True Positive Rate", fontsize=10)
plt.tick_params(axis="both", which="major", labelsize=12)
plt.legend(loc="lower right", fontsize=7, frameon=False)
plt.show()