In [1]:

from sklearn import ensemble
from sklearn import preprocessing

In [2]:

import pandas as pd

In [3]:

train = pd.read_csv('train.csv')

In [4]:

# データ加工
train.Age = train.Age.fillna(train.Age.mean())

le_sex = preprocessing.LabelEncoder()
train.Sex = le_sex.fit_transform(train.Sex)
print train.Sex.head()

le_embarked = preprocessing.LabelEncoder()
train.Embarked = le_embarked.fit_transform(train.Embarked)
print train.Embarked.head()

0    1
1    0
2    0
3    0
4    1
Name: Sex, dtype: int64
0    3
1    1
2    3
3    3
4    3
Name: Embarked, dtype: int64

In [5]:

y = train['Survived']
X = train[['Age', 'Sex', 'Pclass', 'SibSp', 'Parch', 'Fare', 'Embarked']]

In [6]:

rf = ensemble.RandomForestClassifier()

In [7]:

rf.fit(X, y)

Out[7]:

RandomForestClassifier(bootstrap=True, compute_importances=None,
            criterion='gini', max_depth=None, max_features='auto',
            min_density=None, min_samples_leaf=1, min_samples_split=2,
            n_estimators=10, n_jobs=1, oob_score=False, random_state=None,
            verbose=0)

In [8]:

py = rf.predict(X)
print rf.score(X, y)

0.967452300786

In [9]:

pd.crosstab(py, y)

Out[9]:

Survived	0	1
row_0
0	539	19
1	10	323

In [10]:

from sklearn.metrics import confusion_matrix

In [11]:

confusion_matrix(py, y)

Out[11]:

array([[539,  19],
       [ 10, 323]])

In [12]:

from sklearn.metrics import accuracy_score

In [22]:

accuracy_score(py, y)

Out[22]:

0.96745230078563416