Codes for VaR-Copula

Codes (Latex Environment)

\chapter{附件}

\section{Python代码}

编程软件：Python3.8\ jupyter\ notebook

电脑系统：Window\ 10\ 家庭中文版

电脑型号：ASUS\ TUF\ Gaming\ FA506IU

处理器：AMD\ Ryzen\ 7\ 4800H\ with\ Radeon\ Graphics\ 2.90GHz

电脑内存(RAM)：16.0GB

\subsection{收益率边缘分布}

\begin{lstlisting}[language=Python]

import numpy as np

import pandas as pd

import scipy.stats as st

from scipy.stats import t

import matplotlib.pyplot as plt

plt.rcParams['font.sans-serif'] = ['Microsoft YaHei']

sh = pd.read_csv('上证指数历史数据(2).csv')

sz = pd.read_csv('深证成份指数历史数据(2).csv')

hs = pd.read_csv('香港恒生指数历史数据.csv')

def get_pct(data):

    data.index = data['日期']

    data_pct = data ['涨跌幅']

    return data_pct

sh_pct = get_pct(sh)

sz_pct = get_pct(sz)

hs_pct = get_pct(hs)

#边缘密度

def show_edge(data,choose):

    mu = data.mean()#.values

    sigma = data.std()#.values

    num_bins = 100

    x = mu + sigma * np.random.randn(1000000)

    # 正态分布的数据

    data.hist(bins=300,figsize=(16,8),density = True,range=(-0.1,0.1)

                 ,color='b',alpha=0.6,label='密度函数')

    n, bins, patches = plt.hist(x, num_bins, density = True,alpha=0)

    # 拟合曲线

    y = st.norm.pdf(bins, mu, sigma)

    plt.plot(bins, y,'r', alpha = 0.8,linewidth=2.5,label='正态分布')

    #data.plot(kind = 'kde', color = 'g',linewidth=1.5,alpha=1,

                  label = '核密度图')

    t_dimen = 10

    z = np.linspace(t.ppf(0.01, t_dimen,loc=mu,scale=sigma),t.ppf(0.99

                 ,t_dimen,loc=mu,scale=sigma), 1000)

    plt.plot(z, t.pdf(z, t_dimen,loc=mu,scale=sigma), alpha=1.0

                 ,linewidth=2.5, c='gray',label='t-分布')

    scipy_kde=st.gaussian_kde(data)#高斯核密度估计

    X_plot=np.linspace(-0.1,0.1,1000)

    dens=scipy_kde.evaluate(X_plot)

    plt.plot(X_plot,dens,c='g',linewidth=2.5,label='核密度图')

    font = {'size': 15}#设置横纵坐标的名称以及对应字体格式、大小

    plt.xlabel('Expectation',font)

    plt.ylabel('Probability',font)

    plt.title('histogram of normal distribution: $\mu =$' +

                 str(mu)[:6] + ', $\sigma =$'+ str(sigma)[:6],font)

    axes = plt.gca()

    axes.set_xlim([-0.075,0.075])

    #显示题注

    plt.legend(loc=0)

    plt.legend(fontsize=15)#显示图例,设置图例字体大小

    plt.grid(True)

    plt.savefig(str(choose)+'.png')

    plt.show()

show_edge(sh_pct,choose='sh')

show_edge(sz_pct,choose='sz')

show_edge(hs_pct,choose='hs')

time_preiod = ['2015年12月9日','2016年12月9日','2017年12月8日'

        ,'2018年12月7日','2019年12月9日','2020年12月9日']

def get_period_pct(time_preiod):

    n = len(time_preiod)

    S=[]

    for i in range(n-1):

        sh_mean,sh_std = sh_pct[time_preiod[i]:time_preiod[i+1]]

                 .mean(),sh_pct[time_preiod[i]:time_preiod[i+1]].std()

        sz_mean,sz_std = sz_pct[time_preiod[i]:time_preiod[i+1]]

                 .mean(),sz_pct[time_preiod[i]:time_preiod[i+1]].std()

        hs_mean,hs_std = hs_pct[time_preiod[i]:time_preiod[i+1]]

                 .mean(),hs_pct[time_preiod[i]:time_preiod[i+1]].std()

        S.append([sh_mean,sh_std,sz_mean,sz_std,hs_mean,hs_std])

    return pd.DataFrame(S)

get_period_pct(time_preiod)

\end{lstlisting}

\subsection{相关系数及等高线}

\begin{lstlisting}[language=Python]

def merge(datalist):

    n=len(datalist)

    result = pd.DataFrame(datalist[0]).copy()

    result.columns=[0]

    for i in range(1,n):