ranger

Written in optimized C++ with multithreading support, ranger offers significant speed improvements over other implementations, as evidenced by its focus on high-performance for high-dimensional data.

Supports multiple forest types including standard random forests, extremely randomized trees, and quantile regression forests, providing flexibility for various predictive tasks.

Implements Random Survival Forests for time-to-event data modeling, a feature highlighted in the references and not commonly found in other random forest libraries.

Designed for efficiency with high-dimensional datasets, it can handle large numbers of features without performance degradation, as stated in the introduction.