-
Notifications
You must be signed in to change notification settings - Fork 0
/
knn_classifier.py
50 lines (36 loc) · 1.15 KB
/
knn_classifier.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
import pickle
import json
import sklearn
from sklearn.neighbors import KNeighborsClassifier
from sklearn import metrics
import numpy as np
def knn_class(k, X_train, Y_train, X_test, test_ids):
y_pred = {}
knn = KNeighborsClassifier(n_neighbors = k)
knn.fit(X_train, Y_train)
p = knn.predict(X_test)
print(p)
y_pred[test_ids] = knn.predict(X_test)
return y_pred
def get_feats(file):
feats = []
q_ids = []
for key, value in file.items():
q_ids.append(key)
feats.append(value)
return q_ids, feats
def main():
train = pickle.load(open('data/yn_train_jfeats.pkl', 'rb'))
_, X_train = get_feats(train)
print("Train:", len(X_train))
train_gt = pickle.load(open('data/train_yesno_gt.pkl','rb'))
Y_train = list(train_gt.values())
print("YTrain:", len(Y_train))
test = pickle.load(open('data/non_yn_train_jfeats.pkl','rb'))
q_ids, X_test = get_feats(test)
print("Test:", len(X_test))
pred = knn_class(7, X_train, Y_train, X_test, q_ids)
print("Now Dumping!")
pickle.dump(pred,open('nonyesno_preds_knn.pkl', 'wb'))
if __name__ == "__main__":
main()