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
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
| import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
# データセットの読み込み
train_data = pd.read_csv('/kaggle/input/titanic/train.csv')
test_data = pd.read_csv('/kaggle/input/titanic/test.csv')
# train_dataとtest_dataの連結
test_data['Survived'] = np.nan
df = pd.concat([train_data, test_data], ignore_index=True, sort=False)
# ------------ Age ------------
# Age を Pclass, Sex, Parch, SibSp からランダムフォレストで推定
from sklearn.ensemble import RandomForestRegressor
# 推定に使用する項目を指定
age_df = df[['Age', 'Pclass','Sex','Parch','SibSp']]
# ラベル特徴量をワンホットエンコーディング
age_df=pd.get_dummies(age_df)
# 学習データとテストデータに分離し、numpyに変換
known_age = age_df[age_df.Age.notnull()].values
unknown_age = age_df[age_df.Age.isnull()].values
# 学習データをX, yに分離
X = known_age[:, 1:]
y = known_age[:, 0]
# ランダムフォレストで推定モデルを構築
rfr = RandomForestRegressor(random_state=0, n_estimators=100, n_jobs=-1)
rfr.fit(X, y)
# 推定モデルを使って、テストデータのAgeを予測し、補完
predictedAges = rfr.predict(unknown_age[:, 1::])
df.loc[(df.Age.isnull()), 'Age'] = predictedAges
# ------------ Name --------------
# Nameから敬称(Title)を抽出し、グルーピング
df['Title'] = df['Name'].map(lambda x: x.split(', ')[1].split('. ')[0])
df['Title'].replace(['Capt', 'Col', 'Major', 'Dr', 'Rev'], 'Officer', inplace=True)
df['Title'].replace(['Don', 'Sir', 'the Countess', 'Lady', 'Dona'], 'Royalty', inplace=True)
df['Title'].replace(['Mme', 'Ms'], 'Mrs', inplace=True)
df['Title'].replace(['Mlle'], 'Miss', inplace=True)
df['Title'].replace(['Jonkheer'], 'Master', inplace=True)
# ------------ Surname ------------
# NameからSurname(苗字)を抽出
df['Surname'] = df['Name'].map(lambda name:name.split(',')[0].strip())
# 同じSurname(苗字)の出現頻度をカウント(出現回数が2以上なら家族)
df['FamilyGroup'] = df['Surname'].map(df['Surname'].value_counts())
# 家族で16才以下または女性の生存率
Female_Child_Group=df.loc[(df['FamilyGroup']>=2) & ((df['Age']<=16) | (df['Sex']=='female'))]
Female_Child_Group=Female_Child_Group.groupby('Surname')['Survived'].mean()
# 家族で16才超えかつ男性の生存率
Male_Adult_Group=df.loc[(df['FamilyGroup']>=2) & (df['Age']>16) & (df['Sex']=='male')]
Male_Adult_List=Male_Adult_Group.groupby('Surname')['Survived'].mean()
# デッドリストとサバイブリストの作成
Dead_list=set(Female_Child_Group[Female_Child_Group.apply(lambda x:x==0)].index)
Survived_list=set(Male_Adult_List[Male_Adult_List.apply(lambda x:x==1)].index)
# デッドリストとサバイブリストをSex, Age, Title に反映させる
df.loc[(df['Survived'].isnull()) & (df['Surname'].apply(lambda x:x in Dead_list)),\
['Sex','Age','Title']] = ['male',28.0,'Mr']
df.loc[(df['Survived'].isnull()) & (df['Surname'].apply(lambda x:x in Survived_list)),\
['Sex','Age','Title']] = ['female',5.0,'Mrs']
# ----------- Fare -------------
# 欠損値を Embarked='S', Pclass=3 の平均値で補完
fare=df.loc[(df['Embarked'] == 'S') & (df['Pclass'] == 3), 'Fare'].median()
df['Fare']=df['Fare'].fillna(fare)
# ----------- Family -------------
# Family = SibSp + Parch + 1 を特徴量とし、グルーピング
df['Family']=df['SibSp']+df['Parch']+1
df.loc[(df['Family']>=2) & (df['Family']<=4), 'Family_label'] = 2
df.loc[(df['Family']>=5) & (df['Family']<=7) | (df['Family']==1), 'Family_label'] = 1 # == に注意
df.loc[(df['Family']>=8), 'Family_label'] = 0
# ----------- Ticket ----------------
# 同一Ticketナンバーの人が何人いるかを特徴量として抽出
Ticket_Count = dict(df['Ticket'].value_counts())
df['TicketGroup'] = df['Ticket'].map(Ticket_Count)
# 生存率で3つにグルーピング
df.loc[(df['TicketGroup']>=2) & (df['TicketGroup']<=4), 'Ticket_label'] = 2
df.loc[(df['TicketGroup']>=5) & (df['TicketGroup']<=8) | (df['TicketGroup']==1), 'Ticket_label'] = 1
df.loc[(df['TicketGroup']>=11), 'Ticket_label'] = 0
# ------------- Cabin ----------------
# Cabinの先頭文字を特徴量とする(欠損値は U )
df['Cabin'] = df['Cabin'].fillna('Unknown')
df['Cabin_label']=df['Cabin'].str.get(0)
# ---------- Embarked ---------------
# 欠損値をSで補完
df['Embarked'] = df['Embarked'].fillna('S')
# ------------- 前処理 ---------------
# 推定に使用する項目を指定
df = df[['Survived','Pclass','Sex','Age','Fare','Embarked','Title','Family_label','Cabin_label','Ticket_label']]
# ラベル特徴量をワンホットエンコーディング
df = pd.get_dummies(df)
# データセットを trainとtestに分割
train = df[df['Survived'].notnull()]
test = df[df['Survived'].isnull()].drop('Survived',axis=1)
# データフレームをnumpyに変換
X = train.values[:,1:]
y = train.values[:,0]
test_x = test.values
# ----------- 推定モデル構築(Random Forest) ---------------
# from sklearn.feature_selection import SelectKBest
# from sklearn.ensemble import RandomForestClassifier
# from sklearn.pipeline import make_pipeline
# from sklearn.model_selection import cross_validate
# # 採用する特徴量を25個から20個に絞り込む
# select = SelectKBest(k = 20)
# clf = RandomForestClassifier(random_state = 10,
# warm_start = True,
# n_estimators = 26,
# max_depth = 6,
# max_features = 'sqrt')
# # clf = RandomForestClassifier()
# pipeline = make_pipeline(select, clf)
# pipeline.fit(X, y)
# ----------- 推定モデル構築(LightGBM) ---------------
import lightgbm as lgb
from sklearn.metrics import accuracy_score
# LightGBMの分類器をインスタンス化
gbm = lgb.LGBMClassifier(objective='binary', num_leaves=5, reg_alpha=0, reg_lambda=22)
gbm.fit(X, y)
# ----- Submit dataの作成 -------
PassengerId=test_data['PassengerId']
#predictions = pipeline.predict(test_x) # Random Forestで予測
predictions = gbm.predict(test_x) # LightGBMで予測
submission = pd.DataFrame({"PassengerId": PassengerId, "Survived": predictions.astype(np.int32)})
submission.to_csv("result0323.csv", index=False)
|
