RunasSudo
/
scipy-yli
1
Fork 0
Code Issues 1 Pull Requests Releases Activity

Implement more kwargs in yli.regress

This commit is contained in:
RunasSudo 2023-04-21 18:09:58 +10:00
parent aa88239cb1
commit f9a8a5cf01
Signed by: RunasSudo
GPG Key ID: 7234E476BF21C61A
1 changed files with 18 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

24
yli/regress.py
View File

@ -102,6 +102,7 @@ def regress(
*,
nan_policy='warn',
exposure=None,
method=None, maxiter=None, start_params=None,
bool_baselevels=False, exp=None
):
"""
@ -119,6 +120,9 @@ def regress(
:type nan_policy: str
:param exposure: Column in *df* for the exposure variable (numeric, some models only)
:type exposure: str
:param method: See statsmodels *model.fit*
:param maxiter: See statsmodels *model.fit*
:param start_params: See statsmodels *model.fit*
:param bool_baselevels: Show reference categories for boolean independent variables even if reference category is *False*
:type bool_baselevels: bool
:param exp: Report exponentiated parameters rather than raw parameters, default (*None*) is to autodetect based on *model_class*
@ -135,11 +139,19 @@ def regress(
df_ref = weakref.ref(df)
dmatrices, dep_categories = df_to_dmatrices(df, dep, formula, nan_policy)
# Build function call arguments
fit_kwargs = {}
if exposure is not None:
fit_kwargs['exposure'] = exposure
if method is not None:
fit_kwargs['method'] = method
if maxiter is not None:
fit_kwargs['maxiter'] = maxiter
if start_params is not None:
fit_kwargs['start_params'] = start_params
# Fit model
if exposure is None:
result = model_class.fit(dmatrices[0], dmatrices[1])
else:
result = model_class.fit(dmatrices[0], dmatrices[1], exposure=exposure)
result = model_class.fit(dmatrices[0], dmatrices[1], **fit_kwargs)
# Fill in general information
result.df = df_ref
@ -1131,13 +1143,13 @@ class Poisson(RegressionModel):
return 'Poisson'
@classmethod
def fit(cls, data_dep, data_ind, exposure=None):
def fit(cls, data_dep, data_ind, exposure=None, method='newton', maxiter=None, start_params=None):
result = cls()
result.exp = True
result.cov_type = 'nonrobust'
# Perform regression
raw_result = sm.Poisson(endog=data_dep, exog=data_ind, exposure=exposure, missing='raise').fit(disp=False)
raw_result = sm.Poisson(endog=data_dep, exog=data_ind, exposure=exposure, missing='raise').fit(disp=False, method=method, start_params=start_params)
result.dof_model = raw_result.df_model
result.dof_resid = raw_result.df_resid