python 风险控制

喜欢ヅ旅行 2023-06-06 04:55 9阅读 0赞

> 通常交易策略中会融入多个因子协同触发信号，在N日突破择时策略的基础上引入风险管理因子。该因子采用止盈止损机制来管理可能出现的风险，ATR指标则作为止盈止损的基准值。

#### ATR指标的实现 ####

ATR指标的计算分为以下两步：

第一步为计算真实波幅TR。TR=MAX\[(当日最高价-当日最低价)、abs(当日最高价-昨日收盘价)、abs(昨日收盘价-当日最低价)\]，指的是今日振幅、今日最高价与昨日收盘价之间的波幅、昨日收盘价与今日最低价之间的波幅，取这三者之中的最大值。  
第二步对真实波幅TR进行N日移动平均计算。ATR=MA(TR,N)，常用参数N为14日或21日。

df_stockload = web.DataReader("600410.SS", "yahoo", datetime.datetime(2018,10,1), datetime.datetime(2019,4,1)) 
    df_stockload['art14'] = talib.ATR(df_stockload.High.values,df_stockload.Low.values,df_stockload.Close.values,timeperiod=14)#计算ATR14
    df_stockload['art21'] = talib.ATR(df_stockload.High.values,df_stockload.Low.values,df_stockload.Close.values,timeperiod=21)#计算ATR21

得到的ATR指标序列化成图像如下：  
![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80NDU4MDk3Nw_size_16_color_FFFFFF_t_70]

#### 止盈止损的实现 ####

此处将ATR值作为止盈止损的基准值，止盈值设置为n\_win倍的ATR值，止损值设置为n\_loss倍的ATR值，n\_win和n\_loss分别为最大止盈系数和最大止损系数，此处设置最大止盈系数为2，最大止损系数为0.8，倾向于盈利值要大于亏损值。触发止盈止损条件为：

当n\_win*ATR值 > (今日收盘价格 - 买入价格)，触发止盈信号，卖出股票  
当n\_loss*ATR值 > (买入价格 - 今日收盘价格)，触发止损信号，卖出股票

用根据风险因子，控制买入卖出

import pandas_datareader as web
    # 融入风险管理
    #股票数据获取及处理接口
    import talib
    
    
    def GetStockDatApi(stockName=None,stockTimeS=None,stockTimeE=None, N1=15, N2=5, n_loss=0.8, n_win=2):
    
        stockdata = web.DataReader(stockName, "yahoo", stockTimeS, stockTimeE)
    
        stockdata['N1_High'] = stockdata.High.rolling(window=N1).max()#计算最近N1个交易日最高价
        stockdata['N1_High'] = stockdata.N1_High.shift(1)
        expan_max = stockdata.Close.expanding().max()
        stockdata['N1_High'].fillna(value=expan_max,inplace=True)#目前出现过的最大值填充前N1个nan
    
        stockdata['N2_Low'] = stockdata.Low.rolling(window=N2).min()#计算最近N2个交易日最低价
        stockdata['N2_Low'] = stockdata.N2_Low.shift(1)
        expan_min = stockdata.Close.expanding().min()
        stockdata['N2_Low'].fillna(value=expan_min,inplace=True)#目前出现过的最小值填充前N2个nan
    
        stockdata['atr14'] = talib.ATR(stockdata.High.values, stockdata.Low.values, stockdata.Close.values, timeperiod=14)  # 计算ATR14
        buy_price = 0
        for kl_index, today in stockdata.iterrows():
            if today.Close > today.N1_High:
                print('N_day_buy', kl_index, today.Close)
                buy_price = today.Close
                stockdata.loc[kl_index, 'signal'] = 1
            #到达收盘价少于买入价后触发卖出
            elif (buy_price != 0) and (buy_price > today.Close) and ((buy_price - today.Close) > n_loss * today.atr14):
                print('stop_loss_n', kl_index, today.Close, buy_price)
                stockdata.loc[kl_index, 'signal'] = 0
                buy_price = 0
            #到达收盘价多于买入价后触发卖出
            elif (buy_price != 0) and (buy_price < today.Close) and ((today.Close - buy_price) > n_win * today.atr14):
                print('stop_win_n', kl_index, today.Close, buy_price)
                stockdata.loc[kl_index, 'signal'] = 0
                buy_price = 0
            elif today.Close < today.N2_Low:
                print('N_day_sell', kl_index, today.Close, buy_price)
                stockdata.loc[kl_index, 'signal'] = 0
                buy_price = 0
            else:
                pass
    
        stockdata['signal'].fillna(method = 'ffill',inplace = True)
        stockdata['signal'] = stockdata.signal.shift(1)
        stockdata['signal'].fillna(method = 'bfill',inplace = True)
    
        return stockdata

#### 总结 ####

将ATR止盈止损策略作为风险管理因子与N日突破择时策略相融合，将多个策略作为因子作用在一起判断走势，可以从不同的维度保证交易的可靠性，从而避免策略的不确定性所带来的交易上的风险。

[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80NDU4MDk3Nw_size_16_color_FFFFFF_t_70]: /images/20230601/b16b9f828e384383a584da63a44606cd.png