S+LEAF documentation

S+LEAF is an open-source software that provides a flexible noise model with fast and scalable methods. It is largely inspired by the celerite / celerite2 model proposed by [1], [3]. In particular the modeling of gaussian processes is similar, and uses the same semiseparable matrices representation as celerite. S+LEAF extends the celerite model in several ways:

• It allows to account for close to diagonal (LEAF) noises such as instrument calibration errors (see [2]).

• It allows to model simulatenously several time series with the same Gaussian processes and their derivatives (see [4]).

• It provides an efficient implementation of the FENRIR stellar activity model (see [5])

Please cite [2], [4], and/or [5] if you use S+LEAF in a publication.

Installation

Using conda

The S+LEAF package can be installed using conda with the following command:

conda install -c conda-forge spleaf

Using pip

It can also be installed using pip with:

pip install spleaf

Usage

S+LEAF covariance matrices are generated using the spleaf.cov.Cov class. The covariance matrix is modeled as the sum of different components (or terms), which split into two categories: noise terms and kernel terms (gaussian processes). See the API reference for a list of available terms.

The low level implementation of S+LEAF matrices as defined by [2] is available as the spleaf.Spleaf class, but one typically does not need to directly deal with it.

Examples

• Modeling a Gaussian Process together with calibration noise
• Modeling a Gaussian Process over heterogeneous time series

API Reference

spleaf.cov
spleaf.term
spleaf.fenrir
spleaf

Contribute

Everyone is welcome to open issues and/or contribute code via pull-requests. A SWITCH edu-ID account is necessary to sign in to https://gitlab.unige.ch. If you don’t have an account, you can easily create one at https://eduid.ch. Then you can sign in to https://gitlab.unige.ch by selecting “SWITCH edu-ID” as your organisation.

References

[1]

Foreman-Mackey et al., “Fast and Scalable Gaussian Process Modeling with Applications to Astronomical Time Series”, 2017.

[2] (1,2,3)

Delisle, J.-B., Hara, N., and Ségransan, D., “Efficient modeling of correlated noise. II. A flexible noise model with fast and scalable methods”, 2020.

[3]

Gordon, T. A., Agol, E., Foreman-Mackey, D., “A Fast, Two-dimensional Gaussian Process Method Based on Celerite: Applications to Transiting Exoplanet Discovery and Characterization”, 2020.

[4] (1,2)

Delisle, J.-B., Unger, N., Hara, N., and Ségransan, D., “Efficient modeling of correlated noise. III. Scalable methods for jointly modeling several observables’ time series with Gaussian processes”, 2022.

[5] (1,2)

Hara, N., and Delisle, J.-B., “A statistical model of stellar variability. I. FENRIR: a physics-based model of stellar activity, and its fast Gaussian process approximation”, submitted.