LoginSign Up
← Back to all sparks
S

svines

ANALYTICS
Velocity0.0

R package svines by tnagler — release notes from GitHub.

github.com/tnagler/svines ↗

Stationary vine copulas for time series, released in lockstep with the rest of Nagler's vine stack.

vine-copulastime-seriesdependence-modellingrcppr-package
◆Current state
svines fits stationary vine copula models to multivariate time series, extending the rvinecopulib engine with the serial dependence structure that makes vines usable for temporal data. The visible history is three releases carrying one real addition — pseudo-residual computation and logLik support at 0.2.2 — with the rest tracking its C++ dependency.
◆Where it's heading
This package moves when rvinecopulib moves. The 0.2.4 release exists solely to adapt to a new rvinecopulib version, and 0.2.7 carries auto-generated GitHub release notes with no description at all. It shipped on the same day as kde1d 1.1.1, another package from the same maintainer, which is the pattern to watch: changes in the shared C++ layer surface as near-simultaneous releases across the vine family rather than as independent work.
◆Prediction
The next release most plausibly follows another rvinecopulib update rather than adding modelling capability. Two of the three visible entries carry no substantive notes, so this feed will keep underreporting what changed.

◆Recent moves

  1. 1y ago

    svines 0.2.7

    Auto-generated GitHub release notes pointing at a pull request and a compare link, with no description of what changed. Stamped the same day as kde1d 1.1.1 from the same maintainer, consistent with a coordinated pass across the vine packages.

    View source ↗
  2. 1y ago

    Adapted to new rvinecopulib version

    A single-line adaptation to a new rvinecopulib version. Downstream compatibility with the C++ vine engine, invisible to callers.

    View source ↗
  3. 2y ago

    Pseudo residuals and logLik support added

    Adds a function to compute pseudo residuals and gives logLik an nobs value, with logLik.svine_margin exported. Pseudo residuals are what let a fitted stationary vine be checked against the data, so this is the release that made model diagnostics possible.

    View source ↗