cgraph: Computational Graphs

Allows to create, evaluate, and differentiate computational graphs in R. A computational graph is a graph representation of a multivariate function decomposed by its (elementary) operations. Nodes in the graph represent arrays while edges represent dependencies among the arrays. An advantage of expressing a function as a computational graph is that this enables to differentiate the function by automatic differentiation. The 'cgraph' package supports various functions including basic arithmetic, trigonometry functions, and linear algebra functions. It differentiates computational graphs by reverse automatic differentiation. The flexible architecture of the package makes it applicable to solve a variety of problems including local sensitivity analysis, gradient-based optimization, and machine learning.

Version: 2.0.3
Imports: R6
Suggests: Rgraphviz, testthat
Published: 2018-09-24
Author: Ron Triepels
Maintainer: Ron Triepels <dev at>
License: Apache License 2.0
NeedsCompilation: yes
Materials: NEWS
CRAN checks: cgraph results


Reference manual: cgraph.pdf
Package source: cgraph_2.0.3.tar.gz
Windows binaries: r-devel:, r-release:, r-oldrel:
OS X binaries: r-release: cgraph_2.0.2.tgz, r-oldrel: cgraph_2.0.2.tgz
Old sources: cgraph archive


Please use the canonical form to link to this page.