NEWS

---

tidyabc 0.0.1

• ABC
  • auto calibrate kernel epsilon (? and sample size) on ESS in adaptive fits
• Migrate
  • scoring utilities from ggoutbreak
  • use of empirical CDF in ggoutbreak
• Empirical distributions
  • display splines of empirical fits
  • update splines when using empirical as a link function
  • see: (https://en.wikipedia.org/wiki/Copula_(statistics)#Gaussian_copula)
  • and: (https://en.wikipedia.org/wiki/Sigmoid_function)
• dist_fns extensions:
  • convolution
  • skew

• ai4ci r-universe release.
• ABC estimators with rejection, SMC or adaptive sampling
  • distance functions for Wasserstein and RMSE.
  • modern support for parallel simulation execution using furrr
  • posterior visualisation and resampling
  • convergence statistics and visualisation
• Statistic functions wrapper dist_fns S3 class
  • support for distributions as dataframe column
  • empirical, transformed and mixture distributions in unified framework
  • distribution plotting
• Empirical distributions
  • from weighted data or from cumulative probabilities
  • link function support for constraints on support
  • monotonic spline based fits in Q-Q (cdf) or logit-Z (data) space with support for q, p, d, r functions.
• standalone distribution functions (migrated from ggrrr)
  • e.g. re-parametrisations of standard distributions.
  • constrained gamma distribution
  • logit normal distribution
• S3 class for priors
  • formula based constructor
  • print methods
  • search library paths for distributions.
  • constraints