113 lines
4.1 KiB
Python
113 lines
4.1 KiB
Python
import yfinance as yf
|
|
import pandas as pd
|
|
import numpy as np
|
|
from scipy.optimize import minimize
|
|
import matplotlib.pyplot as plt
|
|
import warnings
|
|
warnings.filterwarnings("ignore")
|
|
|
|
|
|
class Portfolio:
|
|
def __init__(self, data_symbol, line=True,sample_num = 10000):
|
|
self.data_symbol = data_symbol
|
|
self.data = self.create_stock_data(self.data_symbol)
|
|
self.length = len(self.data.columns)
|
|
self.sample_num = sample_num
|
|
self.calculate(line=line)
|
|
|
|
# 用雅虎财经API创建Data
|
|
def create_stock_data(self, data_symbol):
|
|
data = pd.DataFrame()
|
|
for symbol in data_symbol:
|
|
try:
|
|
symbol_data = yf.download(symbol, start=start_date, end=end_date)
|
|
data = pd.concat([data, symbol_data['Close']], axis=1)
|
|
except:
|
|
print("请检查股票代码拼写是否有误！")
|
|
data.columns = data_symbol
|
|
return data
|
|
|
|
def calculate(self, line):
|
|
# 初始化参数
|
|
self.all_weights = np.zeros((self.sample_num, len(self.data.columns)))
|
|
self.ret_arr = np.zeros(self.sample_num)
|
|
self.vol_arr = np.zeros(self.sample_num)
|
|
self.sharpe_arr = np.zeros(self.sample_num)
|
|
self.log_ret = np.log(self.data/self.data.shift(1))
|
|
|
|
for x in range(self.sample_num):
|
|
# 随机给不同资产赋予权重
|
|
self.weights = np.array(np.random.random(self.length))
|
|
self.weights = self.weights/np.sum(self.weights)
|
|
self.all_weights[x,:] = self.weights
|
|
|
|
# 计算期望收益
|
|
# 252是美股每年的交易天数
|
|
self.ret_arr[x] = np.sum( (self.log_ret.mean() * self.weights * 252))
|
|
# 计算期望风险
|
|
self.vol_arr[x] = np.sqrt(np.dot(self.weights.T, np.dot(self.log_ret.cov()*252, self.weights)))
|
|
|
|
# 计算夏普比率
|
|
self.sharpe_arr[x] = self.ret_arr[x]/ self.vol_arr[x]
|
|
|
|
self.frontier_x = []
|
|
self.frontier_y = np.linspace(0.15, 0.5, 200)
|
|
|
|
for possible_return in self.frontier_y:
|
|
cons = ({'type':'eq','fun': self.check_sum}, {'type':'eq','fun':lambda w : self.get_ret_vol_sr(w)[0]-possible_return})
|
|
init_guess = np.ones(self.length) / (self.length)
|
|
bounds = ((0,1),)*(self.length)
|
|
self.result = minimize(self.minimize_volatility, init_guess, method='SLSQP', bounds=bounds, constraints=cons)
|
|
self.frontier_x.append(self.result['fun'])
|
|
|
|
def draw(self, line=True):
|
|
plt.figure(figsize=(20,8))
|
|
plt.scatter(self.vol_arr, self.ret_arr, c=self.sharpe_arr, cmap='viridis', s=10)
|
|
plt.colorbar(label='Sharpe Ratio')
|
|
plt.xlabel('Volatility')
|
|
plt.ylabel('Return')
|
|
plt.xticks(np.arange(0, 0.5, step=0.02))
|
|
if(line):
|
|
plt.plot(self.frontier_x, self.frontier_y, 'r--', linewidth=3)
|
|
plt.show()
|
|
|
|
def report(self):
|
|
self.optim_number = self.sharpe_arr.argmax()
|
|
self.optim_weights = self.all_weights[self.optim_number]
|
|
print("基于所选风险资产的模拟最佳组合: ")
|
|
for i in range(self.length):
|
|
print("股票代码：{}, 比例：{:.2f}%".format(self.data.columns[i], self.optim_weights[i]*100))
|
|
self.best_results_arr = self.get_ret_vol_sr( self.optim_weights)
|
|
print("期望年化收益: {:.2f}%, 风险: {:.2f}%".format(self.best_results_arr[0]*100, self.best_results_arr[1]*100, self.best_results_arr[2]))
|
|
|
|
|
|
def get_ret_vol_sr(self, weights):
|
|
weights = np.array(weights)
|
|
ret = np.sum(self.log_ret.mean() * weights) * 252
|
|
vol = np.sqrt(np.dot(weights.T, np.dot(self.log_ret.cov()*252, weights)))
|
|
sr = ret/vol
|
|
return np.array([ret, vol, sr])
|
|
|
|
def neg_sharpe(self, weights):
|
|
return self.get_ret_vol_sr(weights)[2] * -1
|
|
|
|
def check_sum(self, weights):
|
|
#return 0 if sum of the weights is 1
|
|
return np.sum(weights)-1
|
|
|
|
def minimize_volatility(self, weights):
|
|
return self.get_ret_vol_sr(weights)[1]
|
|
|
|
|
|
####根据需要调整下面的参数####
|
|
if __name__ == "__main__":
|
|
# 数据起始日期
|
|
start_date = "2016-01-01"
|
|
end_date = "2021-01-01"
|
|
# 选择需要哪些资产的数据
|
|
data_array = ["600519.SS", "0700.HK","BABA", "AAPL", "MSFT","^GSPC","GOLD", "TSLA"]
|
|
|
|
portfolio = Portfolio(data_array, sample_num=100000)
|
|
portfolio.report()
|
|
portfolio.draw()
|