Introduction

This GitHub repository contains the code used for the draft "Reservoir Computing in R: a Tutorial for Using reservoirnet to Predict Complex Time-Series". The repository is organized as follows:

jss_reproducibility.R

This R script assesses the reproducibility of the code. It utilizes knitr::purl("draft_jss/draft_jss.Rmd") for full reproducibility. To run this script, the following packages are required:

• dplyr
• ggplot2
• patchwork
• reservoirnet

Additionally, the file data/precomputed_results.rds is needed for the last part of the script, containing precomputed results from a computation cluster server.

data

df_obfuscated_epidemio.rds

This file contains the data used for the last use case of the paper.

precomputed_results.rds

Precomputed results from a computation cluster server for the last use case of the paper.

functions

This folder contains several functions used for the last use case of the paper, including data preparation and model fitting.

script

This folder contains all scripts used on the computation server.

find_hyperparameter

Scripts used to find relevant hyperparameters for both elastic-net and reservoir computing (echo state network abbreviated as ESN).

forecast

Scripts used to forecast with elastic-net and reservoir computing after hyperparameters are set. The file forecast_test_eval.R gathers all the results, aggregates them, and prepares the plot used in the article.