Value at Risk Estimation with Neural Networks – Master thesis in cooperation with Royal Institute of Technology (KTH)

William Karlsson Lille and Daniel Saphir from Royal Institute of Technology (KTH) have published a master thesis in collaboration with Scila. Their thesis “Value at Risk Estimation with Neural Networks: A Recurrent Mixture Density Approach” investigates if neural network techniques can be used to improve value at risk estimations. More specifically, the thesis focuses on recurrent neural networks and a mixture density output layer for generating mixture density distributions of future portfolio returns from which value at risk estimations are derived.