queelius/symlik: v0.4.0 - FittedLikelihoodModel API

Breaking Changes This release introduces a new statsmodels-style API for fitted models. Old API (deprecated) mle, iterations = model.mle(data, init) se = model.se(mle, data) New API fit = model.fit(data, init) fit.params['lambda'] # MLE estimates fit.se['lambda'] # Standard errors fit.llf # Log-likelihood fit.aic, fit.bic # Information criteria fit.conf_int() # Confidence intervals fit.wald_test(null) # Hypothesis testing fit.summary() # Formatted results table New Features FittedLikelihoodModel class with lazy-computed properties Properties: params, se/bse, llf, aic, bic, nobs, df_model, converged Methods: conf_int(), cov_params(), wald_test(), score_test(), lr_test(), summary() Works with both LikelihoodModel and ContributionModel Documentation Updated all tutorials and reference docs for new API Live docs: https://queelius.github.io/symlik/ Full Changelog https://github.com/queelius/symlik/compare/v0.3.0...v0.4.0