def main():
df = get_stock(TICKER, START_DATE, END_DATE)
bt = Backtest(
df,
SmaCross,
cash=INIT_CASH,
trade_on_close=False,
exclusive_orders=True
)
stats, heatmap = bt.optimize(
long_term=range(3, MAX_LONG_TERM + 1, TERM_STEP),
short_term=range(2, MAX_LONG_TERM, TERM_STEP),
return_heatmap=True,
constraint=lambda p: p.short_term < p.long_term)
print(stats)
bt.plot()
plot_heatmaps(heatmap, agg='mean', plot_width=2048, filename='heatmap')
sns.set(font='IPAexGothic')
d = heatmap.reset_index().pivot('short_term', 'long_term', 'SQN')
fig, ax = plt.subplots(figsize=(12, 9))
sns.heatmap(d, square=True, cmap='seismic', center=0, ax=ax)
ax.set_title('移動平均期間の組み合わせによるSQN')
ax.invert_yaxis()
ax.grid()
plt.savefig('strategy4.png')
plt.show()
def main():
df = get_stock(TICKER, START_DATE, END_DATE)
bt = Backtest(
df,
RsiStrategy,
cash=INIT_CASH,
trade_on_close=False,
exclusive_orders=True
)
stats, heatmap = bt.optimize(
timeperiod=range(5, 41),
rsi_upper=range(60, 91),
rsi_lower=range(10, 41),
return_heatmap=True)
print(stats)
bt.plot()
plot_heatmaps(heatmap, agg='mean', plot_width=2048, filename='heatmap')
# 3Dヒートマップで表示
d = heatmap.reset_index()
fig = plt.figure()
ax = Axes3D(fig)
ax.scatter(d['timeperiod'], d['rsi_upper'], d['rsi_lower'], c=d['SQN'])
ax.set_xlabel('RSI期間(日)')
ax.set_ylabel('RSI上限')
ax.set_zlabel('RSI下限')
plt.show()
def test_plot_heatmaps(self):
bt = Backtest(GOOG, SmaCross)
stats, heatmap = bt.optimize(fast=range(2, 7, 2),
slow=range(7, 15, 2),
return_heatmap=True)
with _tempfile() as f:
for agg in ('mean', lambda x: np.percentile(x, 75)):
plot_heatmaps(heatmap, agg, filename=f, open_browser=False)
# Preview
plot_heatmaps(heatmap, filename=f)
time.sleep(5)
def main():
df = get_stock('^N225', START_DATE, END_DATE)
bt = Backtest(df,
SmaCross,
cash=INIT_CASH,
trade_on_close=False,
exclusive_orders=True)
stats, heatmap = bt.optimize(
long_term=range(3, MAX_LONG_TERM + 1, TERM_STEP),
short_term=range(2, MAX_LONG_TERM, TERM_STEP),
return_heatmap=True,
constraint=lambda p: p.short_term < p.long_term)
print(stats)
bt.plot()
plot_heatmaps(heatmap, agg='mean', plot_width=2048, filename='heatmap')
# But people have this enormous faculty of vision, used to make judgements on much larger data sets much faster. # Let's plot the whole heatmap by projecting it on two chosen dimensions. # Say we're mostly interested in how parameters `n1` and `n2`, on average, affect the outcome. hm = heatmap.groupby(['n1', 'n2']).mean().unstack() hm # Let's plot this table using the excellent [_Seaborn_](https://seaborn.pydata.org) package: # + # %matplotlib inline import seaborn as sns sns.heatmap(hm[::-1], cmap='viridis') # - # We see that, on average, we obtain the highest result using trend-determining parameters `n1=40` and `n2=60`, # and it's not like other nearby combinations work similarly well — in our particular strategy, this combination really stands out. # # Since our strategy contains several parameters, we might be interested in other relationships between their values. # We can use # [`backtesting.lib.plot_heatmaps()`](https://kernc.github.io/backtesting.py/doc/backtesting/lib.html#backtesting.lib.plot_heatmaps) # function to plot interactive heatmaps of all parameter combinations simultaneously. # + from backtesting.lib import plot_heatmaps plot_heatmaps(heatmap, agg='mean')
