Predicting Pass Receiver in Football Using Distance Based Features

摘要

This paper presents our approach to the football pass prediction challenge of the Machine Learning and Data Mining for Sport Analytics workshop at ECML/PKDD 2018. Our solution uses distance based features to predict the receiver of a pass. We show that our model is able to improve prediction results obtained on a similar dataset. One particularity of our approach is the use of failed passes to improve the predictions.