pyqrse.fittools package

Submodules

pyqrse.fittools.optimizer module

class pyqrse.fittools.optimizer.QRSEFitter(the_model)

Bases: pyqrse.utilities.mixins.HistoryMixin

fit(data=None, params0=None, summary=False, save=True, use_jac=True, weights=None, hist=False, check=False, silent=True, use_hess=False, smart_p0=True, use_sp=True, **kwargs)
Parameters:
Returns:
kld(params=None, target=None)
Parameters:
  • params –
  • target –
Returns:
klmin(target=None, save=True, use_jac=True, **kwargs)
Parameters:
  • target –
  • save –
  • use_jac –
  • kwargs –
Returns:
set_kl_target(target)
Parameters:target –
Returns:
update_model()

pyqrse.fittools.sampling module

class pyqrse.fittools.sampling.QRSESampler(qrse_model, chain_format='df')

Bases: object

sampler doc_string

a_rates

acceptance rates for the sampler

chain
getdiff(parameter1, parameter2)

Get the difference between the chains of two parameters

Parameters:
  • parameter1 – string name for p1 (i.e. ‘t_buy’)
  • parameter2 – string name for p2 (i.e. ‘t_sell’)
Returns:

np.ndarray
init(*args, **kwargs)

updates sampler with recent activity of the QRSEModel() :return:

marg_like
max_like()
max_params
mcmc(N=1000, burn=0, single=False, ptype='corr', s=1.0, update_hess=False, new=False, use_tqdm=True)
Parameters:
  • N –
  • burn –
  • single –
  • ptype –
  • s –
  • update_hess –
  • new –
Returns:
n_errors
next(sample_fun='joint', **kwargs)
plot(per_row=2, figsize=(12, 4), use_latex=True)

plot(self, per_row=2, figsize=(12, 4)): :param per_row: :param figsize: :return:

plotdiff(parameter1, parameter2, kind='hist', use_latex=True, figsize=None, **kwargs)

Quickly view the difference between the chains of two parameters

Parameters:
  • parameter1 – string name for p1 (i.e. ‘t_buy’)
  • parameter2 – string name for p2 (i.e. ‘t_sell’)
  • kind – ‘hist’ for histogram or ‘line’ for time-series
  • use_latex – use latex version of parameter names. default is True
  • figsize – invokes plt.figure(figsize=figsize).
  • kwargs – additional arguments for sns.distplot() and plt.plot()
Returns:
propose_new(params=None, ptype='corr', s=1.0)
set_params()

Module contents