linearmodels.iv.absorbing.AbsorbingRegressor¶

class linearmodels.iv.absorbing.AbsorbingRegressor(*, cat: DataFrame | None = None, cont: DataFrame | None = None, interactions: list[Interaction] | None = None, weights: ndarray[tuple[int, ...], dtype[float64]] | None = None)[source]¶

Construct a sparse matrix of regressors from components.

Parameters:¶

cat: DataFrame | None = None¶: Categorical variables (factors) to absorb.
cont: DataFrame | None = None¶: Continuous variables to absorb.
interactions: list[Interaction] | None = None¶: Included interactions.
weights: ndarray[tuple[int, ...], dtype[float64]] | None = None¶: Weights, if any. The regressors are multiplied by the square root of the weights.

Methods

Properties

`approx_rank`	Return the number of columns of the regressors, an upper bound of their rank.
`has_constant`	Flag indicating whether the regressors have a constant equivalent.
`hash`	Return a hash of the regressors, which does not depend on their order.
`regressors`	Return the sparse matrix of regressors.