makeyourownmaker/mixup: Create Convex Combinations of Pairs of Examples and their Labels for Data Augmentation
The mixup method enlarges training sets using linear interpolations of features and associated labels as described in https://arxiv.org/abs/1710.09412. It produces virtual feature-target pairs from randomly drawn feature-target pairs in the training data. The strength of interpolation is governed by a mixup hyperparameter. The method is straight-forward and data-agnostic. It should result in a reduction of generalisation error and may help to resist memorisation of corrupt labels.
Getting started
Browse package contents
Package details |
|
|---|---|
| Author | makeyourownmaker [aut, cre] |
| Maintainer | makeyourownmaker <makeyourownmaker@gmx.com> |
| License | GPL-2 | file LICENSE |
| Version | 0.0.1 |
| URL | https://github.com/makeyourownmaker/mixup |
| Package repository | View on GitHub |
| Installation |
Install the latest version of this package by entering the following in R:
|
