Flexible procedures to compute local density-based outlier scores for ranking outliers. Both exact and approximate nearest neighbor search can be implemented, while also accommodating multiple neighborhood sizes and four different local density-based methods. It allows for referencing a random subsample of the input data or a user specified reference data set to compute outlier scores against, so both unsupervised and semi-supervised outlier detection can be implemented.
| Version: | 0.1.1 |
| Depends: | R (≥ 3.2.0) |
| Imports: | stats, RANN, mnormt |
| Published: | 2016-12-22 |
| Author: | Kristopher Williams |
| Maintainer: | Kristopher Williams <kristopher.williams83 at gmail.com> |
| License: | GPL-3 |
| URL: | https://github.com/kwilliams83/ldbod |
| NeedsCompilation: | no |
| Materials: | README |
| CRAN checks: | ldbod results |
| Reference manual: | ldbod.pdf |
| Package source: | ldbod_0.1.1.tar.gz |
| Windows binaries: | r-devel: ldbod_0.1.1.zip, r-release: ldbod_0.1.1.zip, r-oldrel: ldbod_0.1.1.zip |
| OS X Mavericks binaries: | r-release: ldbod_0.1.1.tgz, r-oldrel: ldbod_0.1.1.tgz |
| Old sources: | ldbod archive |
Please use the canonical form https://CRAN.R-project.org/package=ldbod to link to this page.