Package: lagseq 0.1.0

lagseq: Modern Lag Sequential Analysis with Tidy Transition Networks

A modern, tidy, pipe-friendly toolkit for lag sequential analysis of categorical event sequences. A single 'lsa()' constructor fits classical, two-cell, bidirectional, and dominance engines through a pluggable registry, with multi-lag analysis and structural-zero constraints, and every result is read through a verb that returns a tidy one-row-per-observation data frame. A confirmatory testing battery quantifies the evidence behind each claim: sequence-level bootstrap and analytic Dirichlet-Multinomial certainty for edge uncertainty, split-half reliability for the whole network, case-drop stability, permutation tests, and permutation- and Bayesian-based group comparison. Fits visualize through a single 'plot()' verb (residual heatmap, transition network, chord, sunburst, and forest views) and interoperate with the 'tna', 'Nestimate', and 'TraMineR' ecosystems, both ingesting their sequence objects and converting to network objects. All numerical methods are implemented from primary literature and cross-validated against published worked examples and base-R primitives.

Authors:Mohammed Saqr [aut, cre, cph]