ustats - R Interface to Python Tools for Computing Higher-Order U-Statistics
Provides an R interface to the Python package 'u-stats' <https://pypi.org/project/u-stats/> for efficient computation of higher-order U-statistics using Einstein summation notation, implementing the methods of Chen, Zhang, and Liu (2025) <doi:10.48550/arXiv.2508.12627>. The package automatically converts R objects to 'NumPy' or 'PyTorch' tensors via 'reticulate' and supports GPU acceleration when 'PyTorch' with 'CUDA' is available. Python dependencies are declared via 'reticulate' and can be installed automatically on first use. Designed for large-scale statistical estimation where numerical stability and performance are critical.
Last updated
einsumstatisticsu-statisticsustats-r
5.08 score 2 stars 1 dependents 6 scripts 268 downloadsHOIF - Higher-Order Influence Function Estimators for the Average Treatment Effect
Implements Higher-Order Influence Function (HOIF) estimators of the Average Treatment Effect (ATE), following Robins et al. (2008) <doi:10.1214/193940307000000527>, Robins et al. (2017) <doi:10.1214/16-AOS1515>, Liu et al. (2017) <doi:10.48550/arXiv.1705.07577> and Liu and Li (2023) <doi:10.48550/arXiv.2302.08097>. Estimators of any order are supported, with optional covariate basis transformations (B-splines, Fourier) and optional K-fold sample splitting (cross-fitting) for improved finite-sample performance. The core higher-order U-statistics are computed exactly via the 'ustats' package, an R interface to the 'Python' package 'u-stats'; the underlying algorithm and its computational complexity are analyzed in Chen, Zhang and Liu (2025) <doi:10.48550/arXiv.2508.12627>. A pure R implementation (up to order 6) is also provided as a fallback that does not require 'Python'.
Last updated
causal-inferencehigher-order-influence-functionhigher-order-u-statisticsu-statisticsustats-r
5.07 score 3 stars 13 scripts 242 downloads