SpectralMM

SpectralMM fits generalized linear and robust regression models with spectral majorization and iterative linear solvers. The Julia frontend defaults to pure Julia. R and Python use a shared C++ implementation; Julia can select it with backend=:cpp.

Start with Installation and Julia examples. The modeling interfaces describe matrix and formula inputs in all three languages. See solvers and diagnostics for numerical controls and post-fit inference for uncertainty estimates and their limitations.

The supported statistical accessors include coef, fitted, residuals, nobs, predict, vcov, stderror, confint, coeftable, deviance, and loglikelihood, where applicable. modelmatrix is not implemented for StatisticalResult; retain the input design if it is needed after fitting.

See the Julia API reference for signatures, keyword options, return values and source links.