-
Notifications
You must be signed in to change notification settings - Fork 1
/
ts_classification_2.py
123 lines (96 loc) · 3.12 KB
/
ts_classification_2.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
import matplotlib.pyplot as plt
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.neighbors import KNeighborsClassifier
from sklearn.preprocessing import LabelBinarizer
from tslearn.preprocessing import TimeSeriesScalerMeanVariance
from music import MusicDB
"""CLASSIFICAZIONE KNN DATASET COMPLETO"""
musi = MusicDB()
print(musi.df.info())
print(musi.feat["enc_genre"].unique())
X = musi.df
y = musi.feat["enc_genre"]
scaler = TimeSeriesScalerMeanVariance()
X = scaler.fit_transform(X).reshape(X.shape[0], X.shape[1])
X_train, X_test, y_train, y_test = train_test_split(
X, y, test_size=0.2, random_state=100, stratify=y
)
def draw_confusion_matrix(Clf, X, y):
titles_options = [
("Confusion matrix, without normalization", None),
("KNN-TimeSeries Confusion matrix", "true"),
]
for title, normalize in titles_options:
disp = plot_confusion_matrix(Clf, X, y, cmap="OrRd", normalize=normalize)
disp.ax_.set_title(title)
plt.show()
"""KNN"""
clf = KNeighborsClassifier(n_neighbors=29, weights="distance", p=2)
clf.fit(X_train, y_train)
# Best parameters: {'n_neighbors': 29, 'p': 2, 'weights': 'distance'}
# 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(8):
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(8):
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("KNN-TimeSeries 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()
"""
Grid Search KNN
print("STA FACENDO LA GRIDSEARCH")
param_list = {
"n_neighbors": list(np.arange(1, 30)),
"weights": ["uniform", "distance"],
"p": [1, 2],
}
# grid search
clf = KNeighborsClassifier()
grid_search = GridSearchCV(clf, param_grid=param_list)
grid_search.fit(X_train, y_train)
# results of the grid search
print("\033[1m" "Results of the grid search" "\033[0m")
print()
print("Best parameters: %s" % grid_search.best_params_)
print("Best estimator: %s" % grid_search.best_estimator_)
print()
print("Best k ('n_neighbors'): %s" % grid_search.best_params_["n_neighbors"])
print()
"""