#coding=gbk
import numpy as np
import pandas as pd
from sklearn.metrics import accuracy_score
import matplotlib.pyplot as plt
import seaborn as sns 
from sklearn.linear_model import LogisticRegression
from sklearn.ensemble import RandomForestClassifier
from sklearn.svm import SVC
from sklearn.model_selection import cross_val_score
from sklearn.metrics import precision_recall_curve
from sklearn.metrics import roc_curve, auc
import matplotlib.pyplot as plt
import warnings
warnings.filterwarnings("ignore")

#数据进一步处理，删去Survived列
train=pd.read_csv("E:\\python1\\123\\Data\\new_train.csv")
test=pd.read_csv("E:\\python1\\123\\Data\\new_test.csv")
gender=pd.read_csv("E:\\python1\\123\\Data\\gender_submission.csv")
sur_train=train.Survived
train.drop(['Survived'], axis=1, inplace =True)
test.drop(['Survived'], axis=1, inplace =True)

#逻辑回归
def Logistic_Regression():
    LRmodel=LogisticRegression()
    LRmodel.fit(train,sur_train)
    LR_prediction=LRmodel.predict(test) 
    # 0.9186602870813397
    print("逻辑回归模型accuracy_score验证准确率：",accuracy_score(gender['Survived'], LR_prediction))
    LR_cross_val_scores = cross_val_score(LRmodel, train, sur_train, cv=5)
    print("逻辑回归模型交叉验证准确率:", LR_cross_val_scores.mean()) 
    return LR_prediction

#随机森林分类器
def Forest():
    Fmodel=RandomForestClassifier(n_estimators=500,max_depth=10,min_samples_split=1.0,min_samples_leaf=1)
    Fmodel.fit(train,sur_train)
    F_prediction=Fmodel.predict(test)
    print("随机森林模型accuracy_score验证准确率：",accuracy_score(gender['Survived'], F_prediction))
    F_cross_val_scores = cross_val_score(Fmodel, train, sur_train, cv=5)
    print("随机森林模型交叉验证准确率:", F_cross_val_scores.mean())
    #0.6363636363636364
    return F_prediction

#向量机模型
def SVM():
    SVMmodel=SVC()
    SVMmodel.fit(train,sur_train)
    SVM_prediction=SVMmodel.predict(test)
    #0.6483253588516746
    print("向量机模型accuracy_score验证准确率：",accuracy_score(gender['Survived'], SVM_prediction))
    SVM_cross_val_scores = cross_val_score(SVMmodel, train, sur_train, cv=5)
    print("向量机模型交叉验证准确率:", SVM_cross_val_scores.mean())
    return SVM_prediction


# 绘制 Precision-Recall 曲线
#PR曲线越靠近右上角，模型性能越好
def PrecisionRecall(LR_prediction,F_prediction,SVM_prediction):
    LR_precision, LR_recall, _ = precision_recall_curve(gender['Survived'], LR_prediction)
    F_precision, F_recall, _ = precision_recall_curve(gender['Survived'], F_prediction)
    SVM_precision, SVM_recall, _ = precision_recall_curve(gender['Survived'], SVM_prediction)
    plt.figure(figsize=(8, 6))
    plt.plot(LR_recall, LR_precision, label='LR', marker='o')
    plt.plot(F_recall, F_precision, label='FOREST', marker='o')
    plt.plot(SVM_recall, SVM_precision, label='SVM', marker='o')
    plt.xlabel('Recall')
    plt.ylabel('Precision')
    plt.title('Precision-Recall 曲线')
    plt.legend()
    plt.grid(True)
    plt.show()

#绘制ROC曲线
def ROC(LR_prediction,F_prediction,SVM_prediction):
    LR_fpr, LR_tpr, _ = roc_curve(gender['Survived'], LR_prediction)
    LR_auc = auc(LR_fpr, LR_tpr)
    F_fpr, F_tpr, _ = roc_curve(gender['Survived'], F_prediction)
    F_auc = auc(F_fpr, F_tpr)
    SVM_fpr, SVM_tpr, _ = roc_curve(gender['Survived'], SVM_prediction)
    SVM_auc = auc(SVM_fpr, SVM_tpr)
    plt.figure(figsize=(8, 6))
    plt.plot(LR_fpr, LR_tpr, color='blue', lw=2, label='LR (AUC = %0.2f)' % LR_auc)
    plt.plot(F_fpr, F_tpr, color='green', lw=2, label='Forest (AUC = %0.2f)' % F_auc)
    plt.plot(SVM_fpr, SVM_tpr, color='red', lw=2, label='SVM (AUC = %0.2f)' % SVM_auc)
    plt.plot([0, 1], [0, 1], color='gray', lw=1, linestyle='--')
    plt.xlim([0.0, 1.0])
    plt.ylim([0.0, 1.05])
    plt.xlabel('False Positive Rate (FPR)')
    plt.ylabel('True Positive Rate (TPR)')
    plt.title('ROC')
    plt.legend(loc="lower right")
    plt.grid(True)
    plt.show()


if __name__ == '__main__':
    LR_prediction=Logistic_Regression()
    F_prediction=Forest()
    SVM_prediction=SVM()
    PrecisionRecall(LR_prediction,F_prediction,SVM_prediction)
    ROC(LR_prediction,F_prediction,SVM_prediction)