sacf 2.0

A selective estimator of the autocorrelation function for non-uniformly sampled timeseries data

Selective Estimator for the Autocorrelation Function

S-ACF: A selective estimator for the autocorrelation function of irregularly sampled time series (credit Lars Kreutzer, c++ implementation by Josh Briegal [e-mail])

Installation

Requirements:

• CMAKE (https://cmake.org) > 3.8.
• C++14

From above top level directory run

pip install ./sacf

in python:

SACF follows Astropy LombScargle implementation:

from sacf import SACF

lag_timeseries, correlations = SACF(timeseries, values, errors=None).autocorrelation()

with options:

sacf.autocorrelation(max_lag=None, lag_resolution=None, selection_function='natural', weight_function='fast', alpha=None)

NOTE: If users specify selection_function="fast", weight_function="fractional_squared" or weight_function="gaussian", a python implementation of the SACF will be invoked which is considerably slower than the default C++ option.

Tests

From root directory run:

tox