GermanyLogistics and transport2018 – ongoingDeutsche Bahn
Machine-learning passenger delay forecasts
Deutsche Bahn generated about 150 million passenger-information forecasts daily
Problem
Passengers needed timely estimates of changing train arrival and departure times.
Approach
DB and T-Systems developed machine-learning forecasts using historical and live operating data. DB distributed predictions through channels including DB Navigator.
Results
Takeaway
Forecast volume described operating scale, rather than prediction accuracy.
Categories
- Size
- Large enterprise
- Industry
- Logistics and transport
- Business function
- Customer service
- Innovation type
- Process innovation
- AI technology
- Predictive analytics and ML
- Adoption
- Partner co-development
- Evidence
- Company's own claim
Sources
Source summaries
DB described machine-learning passenger forecasts. It reported about 150 million predictions daily for about 20,000 journeys.
Telekom described the forecasting system being implemented with Deutsche Bahn in 2018. It explained the use of historical training data and live train positions; deployment benefits were prospective at that point.
Full details are in the original sources.
