Abstract
ODERIM “Outlier Detection/Estimation and mitigation for RIsk Management and control, based on Advanced SSP methods, with a focus on extreme situations”.
The long lasting crisis situation since 2008 is corrupting financial data with an increasing number of extreme events (i.e. outliers). These outliers require being detected, processed and, if possible, anticipated in order to keep acceptable performance while limiting specific risks for either long-term management style or high frequency trading. The objective of the project is to improve and optimize statistical filtering techniques (such as Lq-regularized Kalman, MCMC algorithms, Particle filtering) to detect, estimate and mitigate the outliers that occur in financial data in order to avoid the contamination of the systematic exposures due to idiosyncratic (exogenous) extreme events.