def bootstrap_rules(self, n=10000, **kw):
z = self.forecast_returns(**kw)
a = pd.Series({k: v.shape[0] for k, v in z.items()})
b = (a / a.sum())
sharpes = []
corrs = {}
for k, v in b.iteritems():
for x in range(0, int(round(v * n))):
sample = draw_sample(z[k], 252)
sharpes.append(sharpe(sample).rename(k))
corrs[(k, x)] = (sample.resample('W').sum().corr())
return pd.DataFrame(sharpes), pd.Panel(corrs).mean(axis=0)
def bootstrap_rules(self, n=10000, **kw):
z = self.forecast_returns(**kw)
a = pd.Series({k: v.shape[0] for k, v in z.items()})
b=(a/a.sum())
sharpes = []
corrs = {}
for k, v in b.iteritems():
for x in range(0,int(round(v*n))):
sample = draw_sample(z[k], 252)
sharpes.append(sharpe(sample).rename(k))
corrs[(k, x)] = (sample.resample('W').sum().corr())
return pd.DataFrame(sharpes), pd.Panel(corrs).mean(axis=0)
def sharpe(self):
return sharpe(np.trim_zeros(self.returns().sum(axis=1)))
def gross_sharpe(self):
return sharpe(
np.trim_zeros((self.position_returns() -
self.transaction_returns()).sum(axis=1)))
def sharpe(self):
return sharpe(np.trim_zeros(self.returns().sum(axis=1)))
def gross_sharpe(self):
return sharpe(np.trim_zeros((self.position_returns() - self.transaction_returns()).sum(axis=1)))