TITANIC DATA PREDICTATION
ok, 這是我自己做的titanic建模，主要過(guò)程是：
探索數(shù)據(jù)：
看下數(shù)據(jù)維度，數(shù)據(jù)簡(jiǎn)單統(tǒng)計(jì)
數(shù)據(jù)類型（數(shù)值型，對(duì)象型）、
是否有空值，多少空值
用matplotlib繪圖探索數(shù)據(jù)
數(shù)據(jù)預(yù)處理：
填充空值數(shù)據(jù)
轉(zhuǎn)換分類變量
數(shù)據(jù)標(biāo)準(zhǔn)化，可在建模時(shí)轉(zhuǎn)換
feature很多時(shí)要降維 PCA SVD
數(shù)據(jù)建模：
模型評(píng)估：
分類：precision/recall ,f1 score, pr tradeoff圖
回歸： MAE/ MSE/ R平方
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
%matplotlib inline
import dataset
train = pd.read_csv("E:/figure/titanic/train.csv")
test = pd.read_csv("E:/figure/titanic/test.csv")
train.head()
Paste_Image.png
?Survived: Outcome of survival (0 = No; 1 = Yes)
?Pclass: Socio-economic class (1 = Upper class; 2 = Middle class; 3 = Lower class)
?Name: Name of passenger
?Sex: Sex of the passenger
?Age: Age of the passenger (Some entries contain NaN)
?SibSp: Number of siblings and spouses of the passenger aboard
?Parch: Number of parents and children of the passenger aboard
?Ticket: Ticket number of the passenger
?Fare: Fare paid by the passenger
?Cabin Cabin number of the passenger (Some entries contain NaN)
?Embarked: Port of embarkation of the passenger (C = Cherbourg; Q = Queenstown; S = Southampton)
explore dataset
print(train.shape)
print(test.shape)
(891, 12)
(418, 11)
train.dtypes
PassengerId int64
Survived int64
Pclass int64
Name object
Sex object
Age float64
SibSp int64
Parch int64
Ticket object
Fare float64
Cabin object
Embarked object
dtype: object
test.dtypes
PassengerId int64
Pclass int64
Name object
Sex object
Age float64
SibSp int64
Parch int64
Ticket object
Fare float64
Cabin object
Embarked object
dtype: object
train.describe()
Paste_Image.png
merge training set and testing set
full =pd.merge(train,test,how="outer")
full.info()
<class 'pandas.core.frame.DataFrame'>
Int64Index: 1309 entries, 0 to 1308
Data columns (total 12 columns):
PassengerId 1309 non-null float64
Survived 891 non-null float64
Pclass 1309 non-null float64
Name 1309 non-null object
Sex 1309 non-null object
Age 1046 non-null float64
SibSp 1309 non-null float64
Parch 1309 non-null float64
Ticket 1309 non-null object
Fare 1308 non-null float64
Cabin 295 non-null object
Embarked 1307 non-null object
dtypes: float64(7), object(5)
memory usage: 107.4+ KB
#check missing value
full.isnull().sum()
PassengerId 0
Survived 418
Pclass 0
Name 0
Sex 0
Age 263
SibSp 0
Parch 0
Ticket 0
Fare 1
Cabin 1014
Embarked 2
dtype: int64
print(full["Pclass"].value_counts(),"\n",full["Sex"].value_counts(),"\n",full["Embarked"].value_counts())
3 709
1 323
2 277
Name: Pclass, dtype: int64
male 843
female 466
Name: Sex, dtype: int64
S 914
C 270
Q 123
Name: Embarked, dtype: int64
data preprocessing
#dummy variable
full=pd.get_dummies(full,columns=["Pclass","Embarked","Sex"])
#drop variable
full=full.drop(["Cabin","PassengerId","Name","Ticket"],axis=1)
full.dtypes
Survived float64
Age float64
SibSp float64
Parch float64
Fare float64
Pclass_1.0 float64
Pclass_2.0 float64
Pclass_3.0 float64
Embarked_C float64
Embarked_Q float64
Embarked_S float64
Sex_female float64
Sex_male float64
dtype: object
#replace "Age" missing value to -999
full["Age"].fillna(-999,inplace =True)
full["Age"]=full["Age"].astype(int)
full["Age"].hist(bins=10,range=(0,100))
<matplotlib.axes._subplots.AxesSubplot at 0x7ba3750>
Paste_Image.png
#replace nan value with mean
full["Fare"].loc[full["Fare"].isnull()]=full["Fare"].mean()
C:\Anaconda3\lib\site-packages\pandas\core\indexing.py:117: SettingWithCopyWarning:
A value is trying to be set on a copy of a slice from a DataFrame
See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy
self._setitem_with_indexer(indexer, value)
full.isnull().sum()
Survived 418
Age 0
SibSp 0
Parch 0
Fare 0
Pclass_1.0 0
Pclass_2.0 0
Pclass_3.0 0
Embarked_C 0
Embarked_Q 0
Embarked_S 0
Sex_female 0
Sex_male 0
dtype: int64
trainset =full.iloc[:891,]
testset=full.iloc[891:,]
from sklearn.linear_model import LogisticRegression
from sklearn.svm import SVC
from sklearn.ensemble import RandomForestClassifier
from sklearn.naive_bayes import GaussianNB
from sklearn.metrics import f1_score
from sklearn.metrics import precision_recall_curve
x_train = trainset.drop("Survived",axis=1)
y_train = trainset["Survived"]
x_test = testset.drop("Survived",axis=1)
#logistic regression
logreg =LogisticRegression()
logreg.fit(x_train,y_train)
y_predlog = logreg.predict(x_test)
logreg.score(x_train,y_train)
0.7991021324354658
print f1 score and plot pr curve(it's on trainset)
f1_score(y_train,logreg.predict(x_train))
0.72248062015503878
prob = logreg.predict_proba(x_train)
precision, recall, thresholds = precision_recall_curve(y_train,prob[:,1])
plt.plot(precision,recall)
plt.xlabel("precision")
plt.ylabel("recall")
<matplotlib.text.Text at 0xbdc1f50>
Paste_Image.png
from sklearn.metrics import classification_report
print(classification_report(y_train,logreg.predict(x_train),target_names=["unsurvived","survived"]))
precision recall f1-score support
unsurvived 0.81 0.87 0.84 549
survived 0.77 0.68 0.72 342
avg / total 0.80 0.80 0.80 891
#SVC
svc =SVC()
svc.fit(x_train,y_train)
y_predsvc = svc.predict(x_test)
svc.score(x_train,y_train)
0.88776655443322106
#rrandom forest
rf = RandomForestClassifier()
rf.fit(x_train,y_train)
y_predrf = rf.predict(x_test)
rf.score(x_train,y_train)
0.97081930415263751
**feature importances
importanted = rf.feature_importances_
print(importanted)
plt.bar(range(x_train.shape[1]),importanted)
plt.xticks(range(x_train.shape[1]),tuple(x_train.columns),rotation=60)
[ 0.23063155 0.04861788 0.04546831 0.28729027 0.02962199 0.01930221
0.03710312 0.00726902 0.00953395 0.0125742 0.18180419 0.09078331]
([<matplotlib.axis.XTick at 0xd1caed0>,
<matplotlib.axis.XTick at 0xd1cafb0>,
<matplotlib.axis.XTick at 0xd1c6210>,
<matplotlib.axis.XTick at 0xd201630>,
<matplotlib.axis.XTick at 0xd201bf0>,
<matplotlib.axis.XTick at 0xd2061d0>,
<matplotlib.axis.XTick at 0xd206790>,
<matplotlib.axis.XTick at 0xd206d50>,
<matplotlib.axis.XTick at 0xd209330>,
<matplotlib.axis.XTick at 0xd2098f0>,
<matplotlib.axis.XTick at 0xd209eb0>,
<matplotlib.axis.XTick at 0xd20f490>],
<a list of 12 Text xticklabel objects>)
Paste_Image.png
from the pic, you can see "age", "fare"and "sex" are more important variance.
There must be some relation between fare and Pclass, do that later.
#Gaussian NB
nb = GaussianNB()
nb.fit(x_train,y_train)
y_prednb=nb.predict(x_test)
nb.score(x_train,y_train)
0.78114478114478114
#standard
from sklearn.preprocessing import StandardScaler
std = StandardScaler()
std.fit(x_train)
x_train_std = std.transform(x_train)
x_test_std = std.transform(x_test)
#train on this later
#save prediction as csv
#submission = pd.DataFrame({"PassengerId":test["PassengerId"],"Survived":y_predlog})
#submission.to_csv("titanicpred.csv",index=False)
