HH/410

Speaker: Yicheng Chen (McMaster University)

Title: Machine Learning and Bayesian Inference in Finance

Abstract: The traditional paradigm for application of mathematical finance in practice involves choosing a specific model and calibrating its parameters to market data. However, such produce normally has a heavy computational cost. Apart from that, such a method suffers from some problems. The most important one is that the traditional procedure adopts parameters from outsets and ignores the possibility of other family models might do better when circumstance changes. We implement a neural network for updating the log-likelihood values on option pricing models from different families, then compare it to a Bayesian approach proceed in Rogers' paper.