Prediction Intervals in Forecasting – Hilti

 Overview

Company Description

https://careers.hilti.com/en/about-us

Project Description

The S&OP process starts with the statistical forecast, a point estimate for a given period, which is used as in input for the subsequent S&OP process steps.

However, to obtain a more robust and, thus, resilient supply chain planning, also the volatility of the prediction should be taken into consideration. Thus, not only a point estimate but also a bandwidth with a certain probability of the forecast is of interest. Instead of using then just one-point estimate for the, e.g. capacity planning within  the S&OP, the prediction interval can be used to obtain a more robust planning.

Goals of the Project

Develop and implementation of forecasting methods having at least the same quality as the existing statistical forecast, but which also allows the derivation of prediction intervals. A prediction interval is thereby an estimate of an interval in which a future observation will fall, with a certain probability, given what has already been observed.

Deliverables

• Development and implementation of different forecasting methods to derive prediction intervals
• The forecasting methods should obtain at least the same quality, in terms of used KPIs, as the existing methods.

Essential Student Knowledge

• Supply Chain Management
• Statistics
• Machine Learning

More information: escf@tue.nl