Much of his research involves applications of predictive and prescriptive analytics in the area of credit scoring and consumer credit risk modelling. For example, he has researched advanced statistical or machine learning methods to predict Probability of Default (PD), Loss Given Default (LGD), i.e. the proportion of a loan that a lender is unable to recover if the borrower defaults, credit card balance at default, time to default (using survival analysis), and loan profitability.
Development of large-scale electronic structure methods, based on Density Functional Theory within the ONETEP program (onetep.org)
Development of atomistic and multiscale simulation methods for materials using quantum and classical methods, and machine-learned potentials
Application of these simulation methods to discover advanced materials in technologically relevant problems such as batteries, hydrogen fuel cells and drud optimisation
