psytest.info_criteria module

psytest.info_criteria_functions

This module contains functions to calculate various information criteria which allow us to choose the parameters for the ADF test.

psytest.info_criteria.bic(y: ndarray[tuple[int, ...], dtype[float64]], k: int) → float

Bayes Information Criterion (BIC)

Parameters:

• y (numpy.ndarray) – Fitted error values.

• k (int) – Number of parameters in the model.

Returns:

bic – The BIC value.

Return type:

float

psytest.info_criteria.aic(y: ndarray[tuple[int, ...], dtype[float64]], k: int) → float

Akaike Information Criterion (AIC)

Parameters:

• y (numpy.ndarray) – Fitted error values.

• k (int) – Number of parameters in the model.

Returns:

aic – The AIC value.

Return type:

float

psytest.info_criteria.aicc(y: ndarray[tuple[int, ...], dtype[float64]], k: int) → float

Corrected Akaike Information Criterion (AICc)

Parameters:

• y (numpy.ndarray) – Fitted error values.

• k (int) – Number of parameters in the model.

Returns:

aicc – The AICc value.

Return type:

float

psytest.info_criteria.find_optimal_kmax(y: ndarray[tuple[int, ...], dtype[float64]], klimit: int, criteria: Literal['aic', 'bic', 'aicc'] = 'bic') → int

Find the best model based on the given criteria.

Parameters:

• y (list of numpy.ndarray) – List of m data values for each model.

• klimit (int) – Maximum number of lags to consider.

• criteria ("aic", "bic" or "aicc".) – Information criteria to use for evaluation. Defaults to “bic”.

Raises:

• TypeError – If klimit is not an integer

• ValueError – If klimit is negative or greater than sample size. If criteria is not in (“aic”, “bic”, “aicc”)

Returns:

best_model – The optimal value of KMAX.

Return type:

int