flexcode

Submodules

Package Contents

Classes

FlexCodeModel

class FlexCodeModel(model, max_basis, basis_system='cosine', z_min=None, z_max=None, regression_params={}, custom_model=None)[source]

Bases: object

fit(x_train, z_train, weight=None)

Fits basis function regression models.

Parameters:

  • x_train – a numpy matrix of training covariates.

  • z_train – a numpy array of z values.

  • weight – (optional) a numpy array of weights.

Returns:

None.

Return type:

tune(x_validation, z_validation, bump_threshold_grid=None, sharpen_grid=None, n_grid=1000)

Set tuning parameters to minimize CDE loss

Sets best_basis, bump_delta, and sharpen_alpha values attributes

Parameters:

  • x_validation – a numpy matrix of covariates

  • z_validation – a numpy array of z values

  • bump_threshold_grid – an array of candidate bump threshold values

  • sharpen_grid – an array of candidate sharpen parameter values

  • n_grid – integer, the number of grid points to evaluate

Returns:

None

Return type:

predict_coefs(x_new)
predict(x_new, n_grid)

Predict conditional density estimates on new data

n :param x_new: A numpy matrix of covariates at which to predict
param n_grid:

int, the number of grid points at which to

predict the conditional density :returns: A numpy matrix where each row is a conditional density estimate at the grid points :rtype: numpy matrix

estimate_error(x_test, z_test, n_grid=1000)

Estimates CDE loss on test data

Parameters:

  • x_test – A numpy matrix of covariates

  • z_test – A numpy matrix of z values

  • n_grid – Number of grid points at which to predict the

conditional density :returns: an estimate of the CDE loss :rtype: float