macrosynergy.learning.random_effects#

class RandomEffects(group_col='real_date', fit_intercept=True)[source]#

Bases: BaseEstimator

Random effects model for inference on panel data.

Parameters:

  • group_col (str) – The column of a multi-indexed pandas dataframe to group by, for placing of random effects.

  • fit_intercept (bool) – Whether or not to fit an intercept term.

Notes

A random effects model is a way of attributing variance in a dependent variable to variance in a group of observations (i.e. error variance), in a structured manner. In the context of panel data, we often use this to account for cross-sectional correlations by imposing period-specific effects. However, this model is solely for inference and is not predictive.

fit(X, y)[source]#

Fit the random effects model.

Parameters:

  • X (Pandas DataFrame or Series) – Input feature matrix.

  • y (Pandas DataFrame or Series) – Target values with the same index as X.

Returns:

The fitted estimator.

Return type:

self

check_X_y(X, y)[source]#

Checks for the input data.

Parameters:

  • X (pd.DataFrame or pd.Series) – Input feature matrix

  • y (pd.DataFrame or pd.Series) – Target values with the same index as X.

Returns:

The input data.

Return type:

pd.DataFrame, pd.DataFrame

property pvals#

Compute the p-values for the parameter estimates.

Returns:

The p-values.

Return type:

pd.Series