**spdm** is an R package implementing basic operations in the space of **s**ymmetric **p**ositive-**d**efinite (SPD) **m**atrices, aka covariance matrices. The package supports a number of operations, including:

- Implements several forms of regularized covariance estimation, including sparse estimation of large covariance matrices using the graphical lasso.

- Implements several measures of central tendency (means), including "intrinsic" means on the SPD manifold.

- Implements several measures of distance between SPD matrices.

- Implements several forms of interpolation between two covariance matrices.

- Transport of a tangent vector between tangent spaces.

- Implements clustering, classification, and pca in the space of SPD matrices using kernel functions based on distance measures defined for SPD matrices.

### The package can be installed directly from the Gitlab repository using the `devtools` package with Installation

`devtools::install_git('https://gitlab.com/fmriToolkit/spdm.git')`

Alternately, the repository can be cloned to a local machine and installed through the terminal by running

`R CMD INSTALL spdm`

in the directory containing the folder.