Extends the simple k-nearest neighbors algorithm by incorporating numerous kernel functions and a variety of distance metrics. The package takes advantage of 'RcppArmadillo' to speed up the calculation of distances between observations.
| Version: | 1.0.3 |
| Depends: | R (≥ 2.10.0) |
| Imports: | Rcpp (≥ 0.12.5), OpenImageR |
| LinkingTo: | Rcpp, RcppArmadillo |
| Suggests: | testthat, covr, knitr, rmarkdown, irlba |
| Published: | 2016-12-07 |
| Author: | Lampros Mouselimis |
| Maintainer: | Lampros Mouselimis <mouselimislampros at gmail.com> |
| BugReports: | https://github.com/mlampros/KernelKnn/issues |
| License: | MIT + file LICENSE |
| URL: | https://github.com/mlampros/KernelKnn |
| NeedsCompilation: | yes |
| Materials: | README NEWS |
| CRAN checks: | KernelKnn results |
| Reference manual: | KernelKnn.pdf |
| Vignettes: |
binary classification using the ionosphere data Image classification of the MNIST and CIFAR-10 data using KernelKnn and HOG (histogram of oriented gradients) Regression using the Housing data |
| Package source: | KernelKnn_1.0.3.tar.gz |
| Windows binaries: | r-devel: KernelKnn_1.0.3.zip, r-release: KernelKnn_1.0.3.zip, r-oldrel: not available |
| OS X Mavericks binaries: | r-release: KernelKnn_1.0.3.tgz, r-oldrel: KernelKnn_1.0.3.tgz |
| Old sources: | KernelKnn archive |
Please use the canonical form https://CRAN.R-project.org/package=KernelKnn to link to this page.