bic_penalty

bic_penalty(n_samples: int, n_params: int = 1, additional_cpts: int = 1) → float

Create a Bayesian Information Criterion (BIC) penalty.

The BIC penalty is a constant penalty given by (n_params + additional_cpts) * log(n_samples), where n_samples is the sample size, n_params is the number of parameters per segment in the model across all features, and additional_cpts is the number of additional change point parameters per segment. For change detection, this is 1.

Parameters:

n_samplesint

Sample size.

n_paramsint, default=1

Number of model parameters per segment.

additional_cptsint, default=1

Number of additional change point parameters per segment. For change detection, this is 1.

Returns:

float

The BIC penalty value.

Examples

>>> bic_penalty(100)
9.210340371976184