Changelog

markovchain 0.9.0

CRAN release: 2022-07-01

Added explicit structural-zero support to assessStationarity().
Statistical inference functions now return R-standard htest objects and use the standard print.htest() output, while retaining the legacy dof field for compatibility.
Standardized assessOrder() on the same htest convention, including parameter = c(df = ...) and the standard test output.
Expanded the statistical_analysis section with numerical test results and time-homogeneity examples.
Added published literature examples for non-rejection of the Markov property and for empirical-versus-theoretical transition matrices.
Corrected structural-zero regression tests and the assessStationarity() documentation example used by package checks.
Corrected verifyEmpiricalToTheoretical() to use row totals when constructing expected transition counts for its row-wise multinomial likelihood-ratio test. This fixes a historical use of column totals and changes the Kullback et al. benchmark from 6.551795 to 6.518384, without changing its inferential conclusion.
Added explicit absorbingStates support to markovchainFit() for MLE fits of sequence data with known terminal states.

Fix for numerical issues
Removed deprecated .Dim methods
Add check monotonicity
Fixed a division-by-zero vulnerability in the C++ backend of generatorToTransitionMatrix(). The function now safely handles absorbing states (where the diagonal element is exactly 0) without returning NaNs or crashing.
Optimized memory allocation in the internal C++ function clean_nas() (used by createSequenceMatrix() and other fitting functions). By pre-calculating the exact number of valid elements before array instantiation, the package now avoids continuous memory reallocations, resulting in faster parsing of large sequences containing NA values.

CRAN release: 2026-02-02

Handled R_CHECK_PACKAGES_USED_IN_DEMO