statsmodels.genmod.generalized_estimating_equations.GEE.predict

GEE.predict(params, exog=None, offset=None, exposure=None, linear=False)[source]

Return predicted values for a marginal regression model fit using GEE.

Parameters:
  • params (array-like) – Parameters / coefficients of a marginal regression model.
  • exog (array-like, optional) – Design / exogenous data. If exog is None, model exog is used.
  • offset (array-like, optional) – Offset for exog if provided. If offset is None, model offset is used.
  • exposure (array-like, optional) – Exposure for exog, if exposure is None, model exposure is used. Only allowed if link function is the logarithm.
  • linear (bool) – If True, returns the linear predicted values. If False, returns the value of the inverse of the model’s link function at the linear predicted values.
Returns:
Return type:

An array of fitted values

Notes

Using log(V) as the offset is equivalent to using V as the exposure. If exposure U and offset V are both provided, then log(U) + V is added to the linear predictor.