@@ -22325,3 +22325,36 @@ across a wide array of bioinformatic R packages.")
and clustering large sequence datasets using fast alignment-free k-mer
counting and recursive k-means partitioning.")
(license license:gpl3)))
+
+(define-public r-fastshap
+ (package
+ (name "r-fastshap")
+ (version "0.0.5")
+ (source
+ (origin
+ (method url-fetch)
+ (uri (cran-uri "fastshap" version))
+ (sha256
+ (base32
+ "08f25ib5mry6h8lvj0g3clc9kfl5g2wdd8x8bw455wwmbcm6x5vg"))))
+ (properties `((upstream-name . "fastshap")))
+ (build-system r-build-system)
+ (propagated-inputs
+ `(("r-abind" ,r-abind)
+ ("r-ggplot2" ,r-ggplot2)
+ ("r-gridextra" ,r-gridextra)
+ ("r-matrixstats" ,r-matrixstats)
+ ("r-plyr" ,r-plyr)
+ ("r-rcpp" ,r-rcpp)
+ ("r-rcpparmadillo" ,r-rcpparmadillo)
+ ("r-tibble" ,r-tibble)))
+ (home-page
+ "https://github.com/bgreenwell/fastshap")
+ (synopsis "Fast Approximate Shapley Values")
+ (description
+ "Computes fast (relative to other implementations) approximate
+Shapley values for any supervised learning model. Shapley values help
+to explain the predictions from any black box model using ideas from
+game theory; see Strumbel and Kononenko (2014)
+<doi:10.1007/s10115-013-0679-x> for details.")
+ (license license:gpl2+)))