A collection of functions which fit functional neural network models. In other words, this package will allow users to build deep learning models that have either functional or scalar responses paired with functional and scalar covariates. We implement the theoretical discussion found in Thind, Multani and Cao (2020) <arXiv:2006.09590> through the help of a main fitting and prediction function as well as a number of helper functions to assist with cross-validation, tuning, and the display of estimated functional weights.
| Version: | 1.0 |
| Imports: | keras, tensorflow, fda.usc, fda, ggplot2, ggpubr, caret, pbapply, reshape2, flux, doParallel, foreach, Matrix |
| Suggests: | knitr, rmarkdown |
| Published: | 2020-09-15 |
| Author: | Richard Groenewald [ctb], Barinder Thind [aut, cre, cph], Jiguo Cao [aut], Sidi Wu [ctb] |
| Maintainer: | Barinder Thind <barinder.thi at gmail.com> |
| License: | GPL-3 |
| URL: | https://arxiv.org/abs/2006.09590, https://github.com/b-thi/FuncNN |
| NeedsCompilation: | no |
| Citation: | FuncNN citation info |
| Materials: | README |
| CRAN checks: | FuncNN results |
| Reference manual: | FuncNN.pdf |
| Package source: | FuncNN_1.0.tar.gz |
| Windows binaries: | r-devel: FuncNN_1.0.zip, r-release: FuncNN_1.0.zip, r-oldrel: FuncNN_1.0.zip |
| macOS binaries: | r-release: FuncNN_1.0.tgz, r-oldrel: FuncNN_1.0.tgz |
Please use the canonical form https://CRAN.R-project.org/package=FuncNN to link to this page.