nemos.observation_models.GammaObservations.estimate_scale#

GammaObservations.estimate_scale(y, predicted_rate, dof_resid)[source]#

Estimate the scale of the model based on the GLM residuals.

For \(y \sim \Gamma\) the scale is equal to,

\[\Phi = \frac{\text{Var(y)}}{V(\mu)}\]

with \(V(\mu) = \mu^2\).

Therefore, the scale can be estimated as the ratio of the sample variance to the squared rate.

Parameters:

  • y (Array) – Observed neural activity.

  • predicted_rate (Array) – The predicted rate values. This is not used in the Poisson model for estimating scale, but is retained for compatibility with the abstract method signature.

  • dof_resid (Union[float, Array]) – The DOF of the residuals.

Return type:

Union[float, Array]

Returns:

The scale parameter. If predicted_rate is (n_samples, n_neurons), this method will return a scale for each neuron.