Beta Kernel Process Modeling R package BKP version 0.3.1 Implements the Beta Kernel Process (BKP) for nonparametric modeling of covariate-dependent binomial probabilities, and the Dirichlet Kernel Process (DKP) for categorical or multinomial response data. Scalable global-local approximations are provided through TwinBKP and TwinDKP, using twinning-selected global subsets and local nearest-neighbour updates. Functions are included for model fitting, predictive inference with uncertainty quantification, posterior simulation, and visualization in one- and two-dimensional input spaces. Gaussian, Matern 5/2, Matern 3/2, and Wendland kernels are supported, with hyperparameters selected by multi-start derivative-free optimization. For more details, see Zhao, Qing, and Xu (2025) <doi:10.48550/arXiv.2508.10447>. Software Depends: R (>= 3.5.0) Imports: dirmult, ggplot2, gridExtra, lattice, nloptr, rlang, Rcpp, tgp LinkingTo: Rcpp, RcppArmadillo, nloptr Suggests: knitr, mlbench, testthat (>= 3.0.0) Jiangyan Zhao <zhaojy2017@126.com> Comprehensive R Archive Network (CRAN) Jiangyan Zhao [cre, aut], Kunhai Qing [aut], Jin Xu [aut] GPL (>= 3) 2026-07-13 application/tgz https://CRAN.R-project.org/package=BKP doi:10.32614/CRAN.package.BKP