| setartree-package {setartree} | R Documentation |
Getting started with the setartree package
Description
The setartree is a library containing the implementations of SETAR-Tree and SETAR-Forest which are forecasting-specific tree-based models that are in particular suitable for global time series forecasting.
Details
If you have problems using setartree, find a bug, or have suggestions, please file an issue on github (bugs/suggestions). If that fails, then you can contact the maintainer directly by email.
If you use the package, please cite the following work in your publications:
Godahewa, R., Webb, G. I., Schmidt, D., & Bergmeir, C. (2023). SETAR-Tree: A novel and accurate tree algorithm for global time series forecasting. Machine Learning, 112, 2555-2591. doi: 10.1007/s10994-023-06316-x
Demos for using SETAR-Tree and SETAR-Forest are available. To get a list of them, type:
library(setartree)
demo()
To execute the SETAR-Tree demo, type:
demo(tree_demo)
To execute the SETAR-Forest demo, type:
demo(forest_demo)
To fit a SETAR-Tree model either using a list of time series or an embedded input matrix and labels,
use the function setartree.
To fit a SETAR-Forest model either using a list of time series or an embedded input matrix and labels,
use the function setarforest.
To obtain forecasts from a SETAR-Tree or a SETAR-Forest, use the functions forecast.setartree and forecast.setarforest, respectively.
The setartree package also contains three datasets that can be used to train/test the SETAR-Tree and SETAR-Forest models:
chaotic_logistic_series, web_traffic_train and web_traffic_test.
See the setartree user manual for detailed explanations about the datasets and the parameters taken by each function.
Another nice tool is the forecast package, that can be used to
plot the time series together with the forecasts generated by SETAR-Tree or SETAR-Forest.
Author(s)
Rakshitha Godahewa rakshithagw@gmail.com
Christoph Bergmeir christoph.bergmeir@monash.edu
Daniel Schmidt daniel.schmidt@monash.edu
and Geoffrey Webb geoff.webb@monash.edu
Department of Data Science and AI, Faculty of Information Technology, Monash University, Australia.
https://www.monash.edu/it/dsai
References
Godahewa, R., Webb, G. I., Schmidt, D., & Bergmeir, C. (2023). SETAR-Tree: A novel and accurate tree algorithm for global time series forecasting. Machine Learning, 112, 2555-2591. doi: 10.1007/s10994-023-06316-x