基于机器学习的信用风险评估系统源码+数据集(逻辑回归+svm-分析贷款发放风险).zip

# -*- coding: utf-8 -*- """ Created on Sun Mar 6 14:23:54 2022 Author: Fu Yangyang @Our Final variables: pool=['gender', 'housing', 'income', 'std_age', 'past_bad_credit', 'married', '0', '1', 'edu0', 'e1', 'e2', 'e3'] """ import os,csv import pandas as pd import numpy as np from sklearn.linear_model import LogisticRegression import statsmodels.api as sm from sklearn.metrics import roc_curve, auc import matplotlib.pyplot as plt #from tqdm import tqdm from sklearn import preprocessing from imblearn.over_sampling import SMOTE path = r'C:\Users\hp\Desktop\MFIN7034 Problem set2' #df=pd.read_csv(path+os.sep+'credit_risk.csv') with open(path+os.sep+'credit_risk.csv', mode='r') as csv_file: csv_reader = csv.reader(csv_file, delimiter=',') line = 0 data_temp = [] for row in csv_reader: if line == 0: print("variable names",",".join(row)) var_names = row line=line+1 else: data_temp.append(list(map(float,row))) line=line+1 csv_file.close() data_temp_m = np.asmatrix(data_temp) df = pd.DataFrame(np.array(data_temp), columns=var_names) def draw_curve(fpr,tpr,roc_auc,save_name): ###make a plot of roc curve plt.figure(dpi=150) lw = 2 plt.plot(fpr, tpr, color='darkorange', lw=lw, label='ROC curve (area = %0.2f)' % roc_auc) plt.plot([0, 1], [0, 1], color='navy', lw=lw, linestyle='--') plt.xlim([0.0, 1.0]) plt.ylim([0.0, 1.05]) plt.xlabel('False Positive Rate') plt.ylabel('True Positive Rate') plt.title(save_name) plt.legend(loc="lower right") plt.savefig(path+os.sep+save_name+'.jpg') plt.show() print('Figure was saved to ' + path) #%% #Question1 //Simple Logistic model LR = LogisticRegression() ###simple example: predictors include income and past_bad_credit X=df[['income','past_bad_credit']] y=df['default_label'] ###run logistic regression lr_model = LR.fit(X,y) ###another way to run logistic regression lr_model1 = sm.Logit(y,sm.add_constant(X)).fit() ###get a summary result of lr print(lr_model1.summary()) ###this is a two dimensional vector, prob d=0 and prob d=1, use the second one predicted_prob = lr_model.predict_proba(X) predicted_default_prob= predicted_prob[:,1] ###compute false positive rate and true positive rate using roc_curve function fpr, tpr, _ = roc_curve(y, predicted_default_prob) roc_auc = auc(fpr, tpr) draw_curve(fpr,tpr,roc_auc,'2.1 Receiver operating characteristic example') #%% #Question2 // 2.2 Full Logistic Model LR = LogisticRegression(penalty="l1",solver= 'liblinear',class_weight='balanced',tol=0.008,max_iter=100000) #convert the gender, age and income df['gender']=preprocessing.scale(df['gender']) df['std_age']=preprocessing.scale(df['Age']) df['std_income']=preprocessing.scale(df['income']) #change the job_occupation to dummy df['jo_0'] = pd.get_dummies(df['job_occupation'])[0] df['jo_1'] = pd.get_dummies(df['job_occupation'])[1] ##change the edu to dummy df['edu_0'] = pd.get_dummies(df['edu'])[0] df['edu_1'] = pd.get_dummies(df['edu'])[1] df['edu_2'] = pd.get_dummies(df['edu'])[2] df['edu_3'] = pd.get_dummies(df['edu'])[3] ''' #variables that we have tried #df['dummy_edu']=list(map(lambda x: np.log(x),df['edu'])) #df['gender']=list(map(lambda x: np.log(x),df['gender'])) #df['ln_income']=list(map(lambda x: np.log(x),df['income'])) #df['std_ln_age']=preprocessing.scale(list(map(lambda x: np.log(x),df['Age']))) #df['std_ln_income']=preprocessing.scale(list(map(lambda x: np.log(x),df['income']))) #df['Age'] = df['Age']//10 #df['std_edu']=preprocessing.scale(df['edu']) #Use Exhaustive method to try every combination,but at last, we find the best combination is the pool list now LR = LogisticRegression(penalty="l1",solver= 'liblinear',class_weight='balanced',tol=0.008,max_iter=100000) df2=pd.DataFrame() cbna_list=[] #save variables auc_list=[] #save auc variables=[] #save number of variables for i in tqdm(range(7,len(pool)+1)): #Use up to len（pool） variables for cbna in itertools.combinations(pool, i): X=df[list(cbna)] y=df['default_label'] x_smote, y_smote = smote.fit_resample(X, y) lr_model = LR.fit(x_smote,y_smote) predicted_prob = lr_model.predict_proba(x_smote) predicted_default_prob= predicted_prob[:,1] fpr, tpr, _ = roc_curve(y_smote, predicted_default_prob) roc_auc = auc(fpr, tpr) #save results cbna_list.append(list(cbna)) variables.append(len(list(cbna))) auc_list.append(roc_auc) df2['Varibles']=cbna_list df2['No. of variables used'] = variables df2['auc value'] = auc_list ''' #%% #Continue Question 2 #Choose the combinantion that achieve highest auc pool=['gender', 'housing', 'income', 'std_age', 'past_bad_credit', 'married', 'jo_0', 'jo_1', 'edu_0', 'edu_1', 'edu_2', 'edu_3'] smote = SMOTE()#use smote function to balance our data sample X = df[pool] y=df['default_label'] x_smote, y_smote = smote.fit_resample(X, y) lr_model = LR.fit(x_smote,y_smote) lr_model1 = sm.Logit(y_smote,sm.add_constant(x_smote)).fit() predicted_prob = lr_model.predict_proba(x_smote) predicted_default_prob= predicted_prob[:,1] fpr, tpr, _ = roc_curve(y_smote, predicted_default_prob) roc_auc = auc(fpr, tpr) print(lr_model1.summary()) print('the best combination: ', list(X.columns)) print('used variables: ' , len(X.columns)) print('the auc value: ' , roc_auc) draw_curve(fpr,tpr,roc_auc,'2.2 Full Logistic Model') #%% #Question3 // 2.3 SVM from sklearn.svm import SVC regressor = SVC(C=1.0, cache_size=200, class_weight=None, coef0=0.0, decision_function_shape='ovr', degree=3, gamma='auto', kernel='rbf', max_iter=-1, probability=True, random_state=None, shrinking=True, tol=0.001, verbose=False) X = x_smote y = y_smote regressor.fit(X, y) predicted_prob = regressor.predict_proba(X) predicted_default_prob= predicted_prob[:,1] fpr, tpr, _ = roc_curve(y, predicted_default_prob) roc_auc = auc(fpr, tpr) draw_curve(fpr,tpr,roc_auc,'2.3 SVM') #%% #Question4 // 2.4 Out-of-Sample Test from sklearn.model_selection import train_test_split X_train, X_test, y_train, y_test = train_test_split(X, y, train_size=10000, random_state=1) LR = LogisticRegression() lr_model = LR.fit(X_train,y_train) predicted_prob = lr_model.predict_proba(X_test) predicted_default_prob= predicted_prob[:,1] fpr, tpr, _ = roc_curve(y_test, predicted_default_prob) roc_auc = auc(fpr, tpr) draw_curve(fpr,tpr,roc_auc,'2.4 Out-of-Sample Test')