v1.0.4 - new CRAN release

Minor fx in the documentation

v1.0.3 - new CRAN release

• Tiny adaptation due to new Matrix version 1.4-2
• fix for HTML5 in documentation, plus various typos

v1.0.2 - new CRAN release

• Fix linking problem with new version of nloptR (2.0.0)
• Reference to JSS paper

v1.0.1 - new CRAN release

missSBM 1.0.1 [2021-06-04]

• less conservative tests to avoid random failure in CRAN checks
• tiny improvements in partlyObservedNetwork (less storage)

v1.0.0 - new CRAN release

missSBM 1.0.0 (major release)

• now rely on future_lapply for parallel computing (plan to be set by the user)
• faster model exploration (used to be called 'smoothing'), now integrated by default in estimateMissSBM
• Use sparse Matrices to encode 0 and NAs
• complete rewriting of optimization routines (E and M steps) with C++ armadillo routines
• Better initialization and embedded C++ kmeans implementation
• important bug fix in MAR case
• bug fix in inference on covariates
• bug fixed in blockDyad-sampling
• missSBM::SimpleSBM_fit_missSBM now inherits from from sbm::SimpleSBM rather than sbm::SimpleSBM_fit
• change field '$netMatrix' to '$networkData' to comply with new interface in sbm
• defunct functions estimate, sample and simulate are no longer supported

v0.3.0 - new CRAN release

• changing interface after suggestion from JSS reviewers
• updated documentation
• interfacing with package sbm
  • change estimate to estimateMissSBM
  • change sample to observedNetwork
  • use sbm::sampleSimpleSBM instead of missSBM::simulate
  • export less R6 classes for simplification (internal use only)
• some bug fixes
• updated maintainer ([email protected] -> [email protected])

CRAN release 0.2.0

First release available as a CRAN package

binary SBM with

• MAR (node, dyad) missing values
• NMAR sampling design (double-standard, block-node, block-dyad, etc)
• a possible modelling of external covariates
• a basic vignette