-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathLesson14-Validation.py
40 lines (27 loc) · 1.17 KB
/
Lesson14-Validation.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
#!/usr/bin/python
"""
Starter code for the validation mini-project.
The first step toward building your POI identifier!
Start by loading/formatting the data
After that, it's not our code anymore--it's yours!
"""
import pickle
import sys
sys.path.append("../tools/")
from feature_format import featureFormat, targetFeatureSplit
data_dict = pickle.load(open("../final_project/final_project_dataset.pkl", "rb") )
### first element is our labels, any added elements are predictor
### features. Keep this the same for the mini-project, but you'll
### have a different feature list when you do the final project.
features_list = ["poi", "salary"]
data = featureFormat(data_dict, features_list)
labels, features = targetFeatureSplit(data)
### it's all yours from here forward!
from sklearn.tree import DecisionTreeClassifier
clf = DecisionTreeClassifier()
clf.fit(features, labels)
print(clf.score(features,labels))
from sklearn.model_selection import train_test_split
features_train, features_test, labels_train, labels_test = train_test_split(features, labels, test_size=0.3, random_state=42)
clf.fit(features_train, labels_train)
print(clf.score(features_test, labels_test))