For a K = 2 MBMM fit, plots each subject's \(\log L_{i,\code{classes[1]}} - \log L_{i,\code{classes[2]}}\) (positive values favor the first class's reference bank), grouped and jittered by MAP posterior class – the diagnostic used throughout Do, McCarthy & Rogers to sanity-check that the reference banks separate subjects before any gate/covariate information is used.
Usage
# S3 method for class 'mbmm_fit'
plot(x, ...)