API reference#

This is the full API documentation of the skarf toolbox.

Vector autoregressive (VAR) models#

The skarf.var provides estimators for fitting vector autoregressive (VAR) time series models.

Covariance based VAR models#

Adapt a covariance matrix as a linear VAR model via a polynomial fit.

CovarianceVAR(estimator[, order, lag, ...])

Covariance based VAR model.

Regularized linear VAR models#

Regularized linear VAR models.

LinearVAR(estimator[, order, lag, ...])

Linear VAR model.

Multi-sample VAR models#

Fit a separate VAR model for each sample (e.g. subject) in a dataset.

MultiVAR(estimator[, n_jobs])

Multi-sample VAR model.

Covariance estimators#

Extension covariance estimators following sklearn.covariance.

PySPI covariance estimators#

A scikit-learn compatible covariance estimator interface for PySPI SPIs.

SPI(*args, **kwargs)

Abstract minimal interface for PySPI SPI object.

SPICovariance(spi)

Covariance estimator wrapper around a PySPI SPI estimator.

create_spi(name)

Create an SPI by name.

create_spi_from_config(module_name, fcn[, ...])

Create an SPI by its module name and function (with params).

list_available_spis([subset])

List available PySPI SPIs.

load_spi_config_map([subset, cache, cache_dir])

Extract a mapping of SPI identifiers to configuration dicts.