Changes in version 2.0.1                        

Bug fixes

  - Fixed pred() using point estimate instead of random draw for first
    Monte Carlo sample, breaking the parameter-uncertainty integration
  - Fixed confint() ignoring the parm argument. Now honors parameter
    subsetting by name or index, matching the stats::confint contract.
  - Fixed confint() producing NaN without informative error when
    variance-covariance has negative diagonal (bad Hessian). Now gives
    actionable error message.
  - Fixed combine() giving cryptic "non-conformable arrays" error when
    MLEs have different parameter dimensions. Now validates upfront.
  - Fixed sampler.mle_fit_boot() unreachable code branch (length(x$t)
    == 1 is never true for boot objects; corrected to ncol(x$t) == 1L)
  - Removed dead stud case in confint.mle_fit_boot() switch
  - Fixed stale point() references in roxygen and vignette prose (now
    params())
  - Fixed bias.mle_normal vignette example missing ... in method
    signature

                        Changes in version 2.0.0                        

Breaking changes

  - S3 class renamed from "mle" to "mle_fit" to resolve name collision
    with stats4::mle (S4 class). Subclasses follow: "mle_fit_numerical",
    "mle_fit_boot", "mle_fit_rmap". Constructor function names (mle(),
    mle_numerical(), mle_boot()) are unchanged.

  - Removed aic(), bic(), and loglik_val() generics. Use standard R
    generics AIC(), BIC(), and logLik() instead. logLik() returns a
    proper "logLik" object with df and nobs attributes, so AIC() and
    BIC() work automatically via stats::AIC.default.

  - Added coef.mle_fit() method delegating to params() for standard R
    compatibility.

  - Removed mle_weighted() constructor and "mle_weighted" class. Use
    combine() instead — same inverse-variance weighting with better
    error handling and a variadic API.

Improvements

  - Replaced 1:n patterns with seq_len() throughout to avoid edge-case
    bugs
  - Vectorized PI computation in pred()
  - Modernized package documentation ("_PACKAGE" sentinel)
  - Cleaned up fixing/ directory (removed dead experimental code)
  - Integer coercion for n parameter in rmap()

                        Changes in version 1.2.0                        

New features

  - joint() composes independent MLEs with disjoint parameter sets into
    a joint MLE with block-diagonal covariance structure
  - combine() optimally weights independent MLEs for the same parameter
    via inverse-variance (Fisher information) weighting
  - as_dist() converts MLE objects to their asymptotic normal
    distributions, bridging to the algebraic.dist distribution algebra
  - Distribution methods on MLE objects: density(), cdf(), inv_cdf(),
    sup(), dim(), mean(), conditional()

Documentation

  - Rewrote package Description around "MLE as technology" narrative
  - Rewrote README to lead with the algebra (joint, combine, rmap,
    as_dist)
  - Added "The Algebra of MLEs" vignette demonstrating the full
    composition pipeline

Bug fixes

  - Fixed name-propagation issue in rmap() where c() merged parameter
    names with transformation output names

                        Changes in version 1.1.0                        

  - Add coef() S3 method for base R compatibility (delegates to
    params())
  - Add logLik() S3 method returning proper logLik object with df and
    nobs attributes, enabling automatic AIC() and BIC() support from
    base R
  - Fix rmap() to accept numeric n parameter (previously required
    integer)

                        Changes in version 1.0.0                        

  - Initial CRAN release
  - Core MLE class (mle) with methods for:
      - Parameter extraction (params, nparams)
      - Variance-covariance (vcov, se)
      - Confidence intervals (confint)
      - Model comparison (AIC, logLik)
      - Bias and MSE estimation (bias, mse)
      - Fisher information (observed_fim)
      - Sampling from MLE distribution (sampler)
      - Predictive intervals (pred)
      - Expected values (expectation)
      - Marginal distributions (marginal)
  - Numerical optimization wrapper (mle_numerical) for optim() results
  - Bootstrap MLE (mle_boot) for small samples
  - MLE transformations via invariance property (rmap)
  - Weighted combination of MLEs (mle_weighted)
  - Three vignettes demonstrating usage:
      - Fitting common distributions to a DGP
      - Statistics and characteristics of the MLE
      - Data generating processes