机器学习方法汇总——泰坦尼克号之灾难分析

                                  泰坦尼克号之灾难分析整合

 

背景及方法描述：寒小阳——泰坦尼克号之灾分析

包含方法有：Adaboost，GBDT，LR，RF，SVM，VotingC，xgboost等方法。

下载链接：点击打开链接或https://pan.baidu.com/s/1xF_0QdiDZIi61kfCp07zMA  密码：7eof

文件夹内容包括：

python代码：

import numpy as np
import pandas as pd
from sklearn.preprocessing import StandardScaler
from sklearn.ensemble import RandomForestRegressor, RandomForestClassifier
from sklearn.ensemble import AdaBoostClassifier, GradientBoostingClassifier
from sklearn.ensemble import VotingClassifier
from sklearn.linear_model import LogisticRegression
from sklearn.svm import SVC
from sklearn.model_selection import cross_val_score
import os
#########################################################
import numpy as np
import matplotlib.pyplot as plt
from sklearn.model_selection import learning_curve
from pylab import *
mpl.rcParams['font.sans-serif'] = ['SimHei'] # 若不添加，中文无法在图中显示
import matplotlib
matplotlib.rcParams['axes.unicode_minus']=False # 若不添加，无法在图中显示负号
# 用sklearn的learning_curve得到training_score和cv_score，使用matplotlib画出learning curve
###############################################################################
def mkdir(path):
    folder = os.path.exists(path)

    if not folder:  # 判断是否存在文件夹如果不存在则创建为文件夹
        os.makedirs(path)  # makedirs 创建文件时如果路径不存在会创建这个路径

submissionFilesDir = "submissionFiles"
mkdir(submissionFilesDir)
learningCurvesDir = "learningCurves"
mkdir(learningCurvesDir)

    # step 1. 数据读入及预处理
# root_path = 'Datasets\\Titanic'
# data_train = pd.read_csv('%s/%s' % (root_path, 'train.csv'))

data_train = pd.read_csv('data/train.csv')

data_train.info()
print(data_train.describe())



# step 2. 去除唯一属性特
data_train.drop(['PassengerId', 'Ticket'], axis=1, inplace=True)

print(data_train.head())

# step 3. 类别特征One-Hot编码
data_train['Sex'] = data_train['Sex'].map({'female': 0, 'male': 1}).astype(np.int64) #对年龄进行编码，女性female为0，男性male为1
data_train.loc[data_train.Embarked.isnull(), 'Embarked'] = 'S' # 2个Embarked缺失值直接填充为S
data_train = pd.concat([data_train, pd.get_dummies(data_train.Embarked)], axis=1)
data_train = data_train.rename(columns={'C': 'Cherbourg','Q': 'Queenstown','S': 'Southampton'}) #对Embarkd的几种情况进行编码分类


# step 4. 将名字转换
def replace_name(x):
    if 'Mrs' in x: return 'Mrs'
    elif 'Mr' in x: return 'Mr'
    elif 'Miss' in x: return 'Miss'
    else: return 'Other'

data_train['Name'] = data_train['Name'].map(lambda x:replace_name(x)) # 将data_train中的名字替换
data_train = pd.concat([data_train, pd.get_dummies(data_train.Name)], axis=1) # 在将data_train后添加名字列
data_train = data_train.rename(columns={'Miss': 'Name_Miss','Mr': 'Name_Mr',
                                        'Mrs': 'Name_Mrs','Other': 'Name_Other'})

print(data_train.head())

# step 5. 数值特征标准化
def fun_scale(df_feature):
    np_feature = df_feature.values.reshape(-1,1).astype(np.float64)
    feature_scale = StandardScaler().fit(np_feature)
    feature_scaled = StandardScaler().fit_transform(np_feature, feature_scale)
    return feature_scale, feature_scaled

Pclass_scale, data_train['Pclass_scaled'] = fun_scale(data_train['Pclass'])
Fare_scale, data_train['Fare_scaled'] = fun_scale(data_train['Fare'])
SibSp_scale, data_train['SibSp_scaled'] = fun_scale(data_train['SibSp'])
Parch_scale, data_train['Parch_scaled'] = fun_scale(data_train['Parch'])


print(data_train.head(10))
# step 6. 缺失值补全及相应处理
# 处理Age缺失值并标准化
# 缺失值处理函数
def set_missing_feature(train_for_missingkey, data, info):
    known_feature = train_for_missingkey[train_for_missingkey.Age.notnull()].as_matrix()
    unknown_feature = train_for_missingkey[train_for_missingkey.Age.isnull()].as_matrix()
    y = known_feature[:, 0] # 第1列作为待补全属性
    x = known_feature[:, 1:] # 第2列及之后的属性作为预测属性
    rf = RandomForestRegressor(random_state=0, n_estimators=100)
    rf.fit(x, y)
    print(info, "缺失值预测得分", rf.score(x, y))
    predictage = rf.predict(unknown_feature[:, 1:])
    data.loc[data.Age.isnull(), 'Age'] = predictage
    return data

train_for_missingkey_train = data_train[['Age','Survived','Sex','Name_Miss','Name_Mr','Name_Mrs',
                                         'Name_Other','Fare_scaled','SibSp_scaled','Parch_scaled']]
data_train = set_missing_feature(train_for_missingkey_train, data_train,'Train_Age')
Age_scale, data_train['Age_scaled'] = fun_scale(data_train['Age'])

# 处理Cabin特征
def set_Cabin_type(df):
    df.loc[ (df.Cabin.notnull()), 'Cabin' ] = 1.
    df.loc[ (df.Cabin.isnull()), 'Cabin' ] = 0.
    return df

data_train = set_Cabin_type(data_train)

# 5. 整合数据
train_X = data_train[['Sex','Cabin','Cherbourg','Queenstown','Southampton','Name_Miss','Name_Mr','Name_Mrs','Name_Other',
                      'Pclass_scaled','Fare_scaled','SibSp_scaled','Parch_scaled','Age_scaled']].as_matrix()
train_y = data_train['Survived'].as_matrix()

print(train_X)

# 6. 模型搭建及交叉验证
lr = LogisticRegression(C=1.0, tol=1e-6)
svc = SVC(C=1.1, kernel='rbf', decision_function_shape='ovo')
adaboost = AdaBoostClassifier(n_estimators=490, random_state=0)
randomf = RandomForestClassifier(n_estimators=185, max_depth=5, random_state=0)
gbdt = GradientBoostingClassifier(n_estimators=436, max_depth=2, random_state=0)
VotingC = VotingClassifier(estimators=[('LR',lr),('SVC',svc),('AdaBoost',adaboost),
                                      ('RandomF',randomf),('GBDT',gbdt)])

'''
# 交叉验证部分 #####
param_test = {
        'n_estimators': np.arange(200, 240, 1), 
        'max_depth': np.arange(4, 7, 1), 
        #'min_child_weight': np.arange(1, 6, 2), 
        #'C': (1, 1.1, 1.2, 1.3, 1.4, 1.5, 1.6, 1.7, 1.8, 1.9)
        }

from sklearn.grid_search import GridSearchCV

grid_search = GridSearchCV(estimator=xgbClassifier, param_grid=param_test, scoring='roc_auc', cv=5)
grid_search.fit(train_X,train_y)
grid_search.grid_scores_, grid_search.best_params_, grid_search.best_score_
# 交叉验证部分 #####
'''

# 模型训练及交叉验证，同时绘制学习曲线用以分辨该模型的学习状态，过拟合？欠拟合？
def plot_learning_curve(estimator,name, learningCurvesDir,title, X, y, ylim=None, cv=None, n_jobs=1,
                        train_sizes=np.linspace(.05, 1., 20), verbose=0, plot=True):
    """
    画出data在某模型上的learning curve.
    参数解释
    ----------
    estimator : 你用的分类器。
    name: 你用的分类器名称。
    learningCurvesDir: 保存的路径。
    title : 表格的标题。
    X : 输入的feature，numpy类型
    y : 输入的target vector
    ylim : tuple格式的(ymin, ymax), 设定图像中纵坐标的最低点和最高点
    cv : 做cross-validation的时候，数据分成的份数，其中一份作为cv集，其余n-1份作为training(默认为3份)
    n_jobs : 并行的的任务数(默认1)
    """
    train_sizes, train_scores, test_scores = learning_curve(
        estimator, X, y, cv=cv, n_jobs=n_jobs, train_sizes=train_sizes, verbose=verbose)

    train_scores_mean = np.mean(train_scores, axis=1)
    train_scores_std = np.std(train_scores, axis=1)
    test_scores_mean = np.mean(test_scores, axis=1)
    test_scores_std = np.std(test_scores, axis=1)

    if plot:
        fig = plt.figure()
        plt.title(title)
        if ylim is not None:
            plt.ylim(*ylim)
        plt.xlabel(u"训练样本数")
        plt.ylabel(u"得分")
        plt.gca().invert_yaxis()
        plt.grid()

        plt.fill_between(train_sizes, train_scores_mean - train_scores_std, train_scores_mean + train_scores_std,
                         alpha=0.1, color="b")
        plt.fill_between(train_sizes, test_scores_mean - test_scores_std, test_scores_mean + test_scores_std,
                         alpha=0.1, color="r")
        plt.plot(train_sizes, train_scores_mean, 'o-', color="b", label=u"训练集上得分")
        plt.plot(train_sizes, test_scores_mean, 'o-', color="r", label=u"交叉验证集上得分")

        plt.legend(loc="best")

        plt.draw()
        plt.gca().invert_yaxis()
        plt.show()
        fig.savefig(learningCurvesDir+"/"+name + "的学习曲线" + ".png")

    midpoint = ((train_scores_mean[-1] + train_scores_std[-1]) + (test_scores_mean[-1] - test_scores_std[-1])) / 2
    diff = (train_scores_mean[-1] + train_scores_std[-1]) - (test_scores_mean[-1] - test_scores_std[-1])
    return midpoint, diff

######
classifierlist = [('LR',lr),('SVC',svc),('AdaBoost',adaboost),('RandomF',randomf),('GBDT',gbdt),('VotingC',VotingC)]
for name, classifier in classifierlist:
    # 分类器训练与下一步交叉验证无关，训练是为下面测试集预测使用
    classifier.fit(train_X, train_y)
    plot_learning_curve(classifier,name, learningCurvesDir,(name + "的学习曲线"), train_X, train_y)
    print(name, "Mean_Cross_Val_Score is:", cross_val_score(classifier, train_X, train_y, cv=5, scoring='accuracy').mean(), "\n")

# 7. 测试集处理
data_test = pd.read_csv('data/test.csv')
data_test.drop(['Ticket'], axis=1, inplace=True)

data_test['Sex'] = data_test['Sex'].map({'female': 0, 'male': 1}).astype(np.int64)
data_test = pd.concat([data_test, pd.get_dummies(data_test.Embarked)], axis=1)
data_test = data_test.rename(columns={'C': 'Cherbourg','Q': 'Queenstown','S': 'Southampton'})

data_test['Name'] = data_test['Name'].map(lambda x:replace_name(x))
data_test = pd.concat([data_test, pd.get_dummies(data_test.Name)], axis=1)
data_test = data_test.rename(columns={'Miss': 'Name_Miss','Mr': 'Name_Mr',
                                      'Mrs': 'Name_Mrs','Other': 'Name_Other'})

# 测试集标准化函数
def fun_test_scale(feature_scale, df_feature):
    np_feature = df_feature.values.reshape(-1,1).astype(np.float64)
    feature_scaled = StandardScaler().fit_transform(np_feature, feature_scale)
    return feature_scaled

data_test['Pclass_scaled'] = fun_test_scale(Pclass_scale, data_test['Pclass'])
data_test.loc[data_test.Fare.isnull(),'Fare'] = 0 # 缺失值置为0
data_test['Fare_scaled'] = fun_test_scale(Fare_scale, data_test['Fare'])
data_test['SibSp_scaled'] = fun_test_scale(SibSp_scale, data_test['SibSp'])
data_test['Parch_scaled'] = fun_test_scale(Parch_scale, data_test['Parch'])

# 处理测试集Age缺失值并归一化
train_for_missingkey_test = data_test[['Age','Sex','Name_Miss','Name_Mr','Name_Mrs','Name_Other',
                                       'Fare_scaled','SibSp_scaled','Parch_scaled']]
data_test = set_missing_feature(train_for_missingkey_test, data_test, 'Test_Age')
data_test['Age_scaled'] = fun_test_scale(Age_scale, data_test['Age'])

data_test = set_Cabin_type(data_test)

test_X = data_test[['Sex','Cabin','Cherbourg','Queenstown','Southampton','Name_Miss','Name_Mr','Name_Mrs','Name_Other',
                    'Pclass_scaled','Fare_scaled','SibSp_scaled','Parch_scaled','Age_scaled']].as_matrix()


# print(len(classifierlist))
# print(classifierlist)
# 8. 模型预测
for i in range(len(classifierlist)):
    model = classifierlist[i]  # 选择分类器
    print("Test in %s!" % model[0])
    predictions = model[1].predict(test_X).astype(np.int32)
    result = pd.DataFrame({'PassengerId': data_test['PassengerId'].as_matrix(), 'Survived': predictions})
    result.to_csv(submissionFilesDir+"/"+'Result_with_%s.csv' % model[0], index=False)
print('...\nAll Finish!')


# 9. XGBoost
import xgboost as xgb
from sklearn.model_selection import train_test_split

x_train, x_valid, y_train, y_valid = train_test_split(train_X, train_y, test_size=0.1, random_state=0)
print(x_train.shape, x_valid.shape)

xgbClassifier = xgb.XGBClassifier(learning_rate = 0.1,
                                  n_estimators= 234,
                                  max_depth= 6,
                                  min_child_weight= 5,
                                  gamma=0,
                                  subsample=0.8,
                                  colsample_bytree=0.8,
                                  objective= 'binary:logistic',
                                  scale_pos_weight=1)

xgbClassifier.fit(train_X, train_y)
plot_learning_curve(xgbClassifier,"xgboost", learningCurvesDir, ("xgboost" + "的学习曲线"), train_X, train_y) #绘制xgbboost的学习曲线
xgbpred_test = xgbClassifier.predict(test_X).astype(np.int32)
print("Test in xgboost!")
result = pd.DataFrame({'PassengerId':data_test['PassengerId'].as_matrix(), 'Survived':xgbpred_test})
result.to_csv(submissionFilesDir+"/"+'Result_with_%s.csv' % 'XGBoost', index=False)

参考网址：

1、https://blog.csdn.net/han_xiaoyang/article/details/49797143