This project is meant to develop time series forecasting models that would help a restaurant chain better manage their stocks. By predicting the rate at which locations consume basic product the company can optimise stockings and minimise food waste.
As these timeseries have very little noise and patterns are clearly visible when plotting them. We have decided to implement traditional methods of statistical learning, i.e ARIMA. This worked really well and we decreased food waste by 50%
- ARIMA
- R