Changelog

Source: NEWS.md

celecx 0.1.0

• Gaussian process regression learners:
  • LearnerRegrDeepGP / "regr.deepgp" (deepgp package)
  • LearnerRegrGPFit / "regr.gpfit" (GPfit package)
  • LearnerRegrHetGP / "regr.hetgp" (hetGP package)
  • LearnerRegrTGP / "regr.tgp" (tgp package)
• SE estimation learner wrappers:
  • LearnerRegrBootstrapSE / "regr.bootstrap_se.*": bootstrap SE estimation
  • LearnerRegrQuantileSE / "regr.quantile_se.*": quantile prediction to SE conversion
• Learning curve extrapolation objects
  • TaskLCE: Regression / time series task that keeps extra meta-information relevant for learning curve extrapolation
  • ResamplingLCEHoldout, ResamplingLCECV: holdout split and rolling / expanding window cross-validation
  • MeasureLCECoverage, MeasureLCEWinkler: LCE prediction interval coverage and Winkler score (experimental)
• Optimization objectives:
  • ObjectiveDataset: pre-evaluated dataset objective
  • ObjectiveLearner: learner-based objective