Prediction of Failures in the Air Pressure System of Scania Trucks Using a Random Forest and Feature Engineering

This paper demonstrates an approach in data analysis to minimize overall maintenance costs for the air pressure system of Scania trucks. Feature creation on histograms was used. Randomly chosen subsets of attributes were then evaluated to generate an order and a final subset of features. Finally, a Random Forest was applied and fine-tuned. The results clearly show that data analysis in the field is beneficial and improves upon the naive approaches of checking every truck or no truck until failure.