def markowitz(self):
if self.verbose:
print('Optimizing portfolio...linear algebra is fun!!')
mk = mojo.markowitz(self.portfolio)
mk = mk.reshape(mk.shape[0], 1)
#self.optimal = mk
self.optimal = mojo.DataFrame(mk.T,
columns=self.portfolio.columns.tolist())
if self.verbose:
print(
'Master using it and you can have this... Portfolio optimized.'
)
print(self.optimal.head())
def simulate(self, threshold=0.2, days=100, **options):
on = options.pop('on', 'return')
exclude = options.pop('exclude', [])
iterations = options.pop('iterations', 10000)
columns = list(self.portfolio.columns)
for item in exclude:
try:
columns.remove(item)
except:
continue
reduced_portfolio = self.portfolio[columns]
probabilities = []
if self.verbose:
print('Simulating...')
print('Reticulatiing splines...')
for column in reduced_portfolio.columns:
out_dict = {'TKR': column}
sim = mojo.simulate(self.portfolio[column], days, iterations)
prob = mojo.probs_predicted(sim, threshold, on=on)
breakeven = mojo.probs_predicted(sim, 0.0, on=on)
roi = mojo.ROI(sim)
EV = mojo.expected_value(sim)
out_dict.update(
{'Prob. ' + str(100 * threshold) + ' % Return': prob})
out_dict.update({'Breakeven': breakeven})
out_dict.update({'ROI': roi})
out_dict.update({'Expected Value': EV})
probabilities.append(out_dict)
simulation = mojo.DataFrame(probabilities)
simulation = simulation.set_index('TKR')
if self.verbose:
print('Simulation complete.')
print(simulation)
return simulation
def NCO(self, **options):
mu = options.pop('mu', None)
maxNumClusters = options.pop('max_clusters', 10)
if maxNumClusters >= len(self.portfolio.columns):
maxNumClusters = len(self.portfolio.columns) - 1
if self.verbose:
print('Optimizing portfolio...linear algebra is fun!!')
nco = mojo.optPort_nco(self.cov, mu=mu, maxNumClusters=maxNumClusters)
self.nco = mojo.DataFrame(nco.T,
columns=self.portfolio.columns.tolist())
if self.verbose:
print(
'Master using it and you can have this... Portfolio optimized.'
)
print(self.nco.head())
def trend(self, **options):
span = options.pop('span', [3, 10, 1])
exclude = options.pop('exclude', [])
output = mojo.DataFrame()
columns = list(self.portfolio.columns)
for item in exclude:
try:
columns.remove(item)
except:
continue
reduced_portfolio = self.portfolio[columns]
if len(self.portfolio.columns) > 0:
if self.verbose:
print('Reticulating splines...')
if not isinstance(span, list):
print('Error: Span must be a list')
print('Setting span to default: [3, 10, 1]')
span = [3, 10, 1]
df_list = []
for column in reduced_portfolio.columns:
out_trend = mojo.GetBinsFromTrend(self.portfolio[column],
column,
span=span)
df_list.append(out_trend)
output = mojo.merge_stocks(df_list)
else:
print('Grumble, Grumble')
print('Portfolio requires at least one item.')
return output
def hurst(self, **options):
if self.verbose:
print('Calculating Hurst Exponents...')
lag1 = options.pop('lag1', 2)
lag2 = options.pop('lag2', 20)
H = []
for column in self.portfolio.columns:
h0 = mojo.hurst(self.portfolio[column], lag1=lag1, lag2=lag2)
H.append(h0)
H = mojo.np.array(H)
H = H.reshape(H.shape[0], 1)
self.H = mojo.DataFrame(H.T, columns=self.portfolio.columns.tolist())
if self.verbose:
print('Hurst Exponents.')
print(self.H.head())