Changelog
Source:NEWS.md
arf 0.2.2
- Faster and vectorized conditional sampling
- Use min.bucket argument from ranger to avoid pruning if possible
- Option to sample NAs in generated data if original data contains NAs
- Stepsize in forge() to reduce memory usage
- Option for local and global finite bounds
arf 0.2.0
CRAN release: 2024-01-24
- Vectorized adversarial resampling
- Speed boost for compiling into a probabilistic circuit
- Conditional densities and sampling
- Bayesian solution for invariant continuous data within leaf nodes
- New function for computing (conditional) expectations
- Options for missing data