This project uses a Universal Differential Equation (UDE) framework to forecast customer behavior that follows a hypothetical framework. Specifically, this project examines the relative efficacy of traditional, deep learning and UDE methods in forecasting consumer behavior using data from the M5 Forecasting competition where the goal was to provide confidence intervals around consumer demand predictions.
This project uses Julia 1.8.5. juliaup is the recommeneded manager for Julia installations. To install:
- Clone the repository with
git clone https://github.com/DDoyle1066/UDE_Customer_Behavior - Activate the environment and install packages with:
julia> ]
(@v1.8) pkg> activate .
(UDE_Customer_Behavior) pkg> instantiate