R Package	Brief description
e1071	This package provides R users with access to the very popular library of SVM tools and functions, libsvm, written by [c@1392]. Common kernels available include linear, radial, polynomial, and sigmoid and the SVM functions support both binary classification and regression applications. Multicategory outcomes can be used but only via a one-against-one approach in which SVMs are computed for each of the possible pairs of categories in the outcome. The package also includes functions for tuning support vector machines through cross-validation. https://cran.r-project.org/package=e1071
kernlab	This package offers a broad collection of kernel functions, many of which are not available in the e1071 package including hyperbolic tangent and LaPlacian among others. SVM functions support classification (binary and multicategory classes) and regression options. https://cran.r-project.org/package=kernlab
caret	The caret package provides a consistent framework for model calibration, cross-validation, and parameter tuning for SVMs and many other machine learning techniques. The actual SVMs are created by functions from the e1071 package for linear kernels and the kernlab package for all others. It also has the nice feature that it will automatically detect if a parallel environment has been initialized and then divide cross-validation across available computing threads, potentially greatly reducing computing time. https://cran.r-project.org/package=caret