Sign-Language-Interpreter/train_classifier.py at main · itsnotcloud/Sign-Language-Interpreter · GitHub

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
import pickle
import sklearn
from sklearn.ensemble import RandomForestClassifier
from sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score
import numpy as np


data_dict = pickle.load(open('./onehand.pickle', 'rb'))
data_dict2 = pickle.load(open('./twohand.pickle', 'rb'))
data_one = np.asarray(data_dict['data'])
data_two = np.asarray(data_dict2['data'])
labels1 = np.asarray(data_dict['labels'])
labels2 = np.asarray(data_dict2['labels'])


x_train, x_test, y_train, y_test, = train_test_split(data_one, labels1, test_size=0.2, shuffle=True, stratify=labels1)
x_train2, x_test2, y_train2, y_test2, = train_test_split(data_two, labels2, test_size=0.2, shuffle=True, stratify=labels2)

model1 = RandomForestClassifier()
model2 = RandomForestClassifier()

model1.fit(x_train, y_train)
model2.fit(x_train2, y_train2)

y_predict = model1.predict(x_test)
y_predict2 = model2.predict(x_test2)

score1 = accuracy_score(y_predict, y_test)
score2= accuracy_score(y_predict2, y_test2)

print('{}% of samples were classified correctly for model1 !'.format(score1 * 100))
print('{}% of samples were classified correctly for model2 !'.format(score2 * 100))

f = open('model1.p', 'wb')
pickle.dump({'model1': model1}, f)
f.close()
f = open('model2.p', 'wb')
pickle.dump({'model2': model2}, f)
f.close()