NEWS

---

missSBM 1.1.0 (2026-07-21)

Major changes

• Breaking: estimateMissSBM()'s control must now be built with the new missSBM_param() (one named, defaulted argument per option) instead of a raw list(...), which now errors with a message pointing at the replacement.
• missSBM_fit now exposes split(), merge(), candidates_split() and candidates_merge() as instance methods, previously inlined in missSBM_collection's exploration logic; same search algorithm, now independently testable.
• new polish(control) (missSBM_fit and missSBM_collection): node-swap refinement after VEM convergence, fixing individually misclassified nodes that split()/merge() cannot reach.
• requesting more blocks than a network actually supports can make VEM collapse one or more classes; this used to be silent, and split/merge exploration's own repair of it could silently corrupt vBlocks's bookkeeping (duplicated/missing entries, non-smooth ICL/ELBO in plot()). Now fixed and made visible: missSBM_fit$repair(control) recovers a collapsed fit ; new occupiedBlocks/ degenerate fields expose any remaining collapse; bestModel skips degenerate models when possible and plot() marks them with a distinct point shape; estimateMissSBM()'s new stopOnDegenerate/maxConsecutiveDegenerate controls (default TRUE/2) stop forward exploration from growing further into a persistently unsupported range.
• new estimate_chain() (missSBM_collection) / warmChain control (default FALSE, opt-in): initializes each model by splitting the already-converged, smaller neighbor instead of an independent cold clustering, substantially reducing collapse in practice, at the cost of fitting vBlocks sequentially rather than in parallel.
• replace the NLopt/CCSAQ optimizer for the covariate connectivity parameters with a builtin Newton-Raphson solver (the M-step objective is concave, so it converges reliably in a handful of iterations); nloptr is no longer a dependency
• fix partlyObservedNetwork$imputation(): the fill value for missing dyads was biased low by including not-yet-imputed entries in its own computation. Known side effect: doubleStandardSampling_fit's initial psi bootstrap can degenerate when the fill value exactly matches the empirical observed edge rate (refined away by the VEM loop that follows)

Minor changes

• profiled estimateMissSBM() and fixed several hot-path inefficiencies: avoidable dense * sparse promotion overhead, a redundant recomputation of the dense imputed-network matrix, and an ifelse() evaluating both of its branches unconditionally. Roughly 3x faster on our benchmark, bit-identical results
• estimateMissSBM() now sorts/de-duplicates vBlocks if needed (with a warning), since exploration and chaining both assume it is strictly increasing.
• estimateMissSBM()'s control drops the exploration field: nobody used "forward"/ "backward" alone at that level, only "both"/"none", so iterates now doubles as the on/off switch (0 disables exploration entirely). missSBM_collection$explore() no longer reads a stored exploration control field either; its direction argument (default "both") is now a plain per-call parameter, still available for advanced/internal use.
• cap the number of merge candidates tried during backward exploration (control$maxMergeCandidates, default 30) instead of always trying every pair
• fix a crash and a closed-form algebra bug in the "degree" sampling design; parameter recovery for this design can still be biased under heavy missingness (known limitation)
• fix a consistency bug in missSBM_fit$doVEM()'s step-back: only the SBM was restored, not the sampling model or the current imputation
• speed up getCovarArray() and kmeans_missSBM()'s seeding
• remove unused src/utils.h
• missSBM_fit's polish(), repair(), candidates_split() and candidates_merge() gain the same default control as doVEM(), so they can be called standalone without building a full control list first; missSBM_collection's calls (which always pass an explicit control list) are unaffected ; missSBM_collection's estimate()/polish()/explore() now share a control list stored at construction, no longer requiring control on every call.
• Factored missSBM_collection's explore_forward()/explore_backward() into a shared private explore_direction(), and missSBM_fit's candidates_split()/candidates_merge()'s trial-fit loop into a shared private trial_fit_candidates(); verified bit-identical output on the same fixed-seed scenario used throughout this changelog to diagnose VEM collapse

missSBM 1.0.5 (2025-03-13)

• minor fix to comply with nlopt version 2.9.1 (NOCEDAL)

missSBM 1.0.4 (2023-10-24)

• minor fix in package documentation due to evolution of roxygen2 7.0.0 https://github.com/r-lib/roxygen2/issues/1491.

missSBM 1.0.3 (2022-08-23)

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

missSBM 1.0.2 (2022-02-01)

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

missSBM 1.0.1 (2021-06-04)

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

missSBM 1.0.0 (2021-05-25)

• 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

missSBM 0.3.0 (2020-11-18)

• 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])

missSBM 0.2.2

• unexporting sampledNetwork, only use internally
  • merging prepare_data with estimate
  • enhanced documentation
  • moving ownership to großBM

missSBM 0.2.1 (2019-09-16)

• added S3 methods for missSBM_fit, SBM_fit

missSBM 0.2.0 (2019-06-08)

significant changes:

• decent vignette
  • faster tests
  • many bugs corrected

missSBM 0.1.0-9000

• Added a NEWS.md file to track changes to the package.