聚宽源码50

原文策略源码如下：
#阻力支撑相对强度（RSRS）指标择时策略

导入函数库

from jqdata import *
import pandas as pd
import numpy as np
from sklearn import linear_model
from numpy import mean, std

初始化函数，设定要操作的股票、基准等等

def initialize(context):
# 定义一个全局变量, 保存要操作的股票
context.stock_id=‘510300.XSHG’
context.beat_list=[]
context.count=0
context.z_beta_r2_list=[]
context.n_value=18
context.m_value=400
# 设定沪深300作为基准
set_benchmark(‘000300.XSHG’)
# 开启动态复权模式(真实价格)
set_option(‘use_real_price’, True)

def before_trading_start(context):
advance_high_array=attribute_history(context.stock_id,400,‘60m’,‘high’)
advance_low_array=attribute_history(context.stock_id,400,‘60m’,‘low’)

for i in range(context.n_value,context.m_value):
    
    sliding_windows_high_array=advance_high_array[i-18:i]
    sliding_windows_low_array=advance_low_array[i-18:i]

    reg = linear_model.LinearRegression()
    reg.fit (sliding_windows_low_array,sliding_windows_high_array)
    beat=reg.coef_
    context.beat_list.append(beat)

每个单位时间(如果按天回测,则每天调用一次,如果按分钟,则每分钟调用一次)调用一次

def handle_data(context,data):

context.count=context.count+1
if context.count % 60==0:
    high_array=attribute_history(context.stock_id,18,'60m','high')
    low_array=attribute_history(context.stock_id,18,'60m','low')
    reg = linear_model.LinearRegression()
    reg.fit(low_array,high_array)
    beat=reg.coef_
    R_sq=reg.score(low_array,high_array)

    beat_z=R_sq*(beat-mean(context.beat_list))/std(context.beat_list)

    context.z_beta_r2_list.append(beat_z)
    log.info("context.z_beta_r2_list",context.z_beta_r2_list)
    cash = context.portfolio.available_cash
    print(cash)
    print("beat",beat)
    print("R_sq",R_sq)
    print("beat_z",beat_z)
    record(name=beat_z)
    if len(context.z_beta_r2_list)>2:
        if context.z_beta_r2_list[-2]<0.7 and context.z_beta_r2_list[-1]>0.7:
            log.info(" 买入 %s" % (context.stock_id))
            order_value(context.stock_id,cash)
        if context.z_beta_r2_list[-2]>-0.7 and context.z_beta_r2_list[-1]<-0.7:
            log.info(" 卖出 %s" % (context.stock_id))
            order_target_value(context.stock_id, 0)
# record(context.z_beta_r2_list)