fide provides a feature-based forecasting method for intermittent demand proposed by Li Li, Yanfei Kang, Fotios Petropoulos, and Feng Li. The package aims to facilitate reproducing the results of our paper, and can also be applied to other intermittent demand forecasting problems.
You can install fide from github with:
devtools::install_github("lily940703/fide")
library(M4metalearning)
library(tsintermittent)
Simulated data and a real dataset are used in this paper, please see this page for details.
An example of using the package based on simulated data is shown on this page.
- Li, Li, Yanfei Kang, Fotios Petropoulos, and Feng Li. 2022. "Feature-based Intermittent Demand Forecast Combinations: Bias, Accuracy and Inventory Implications." Working Paper.