The Extreme Solution
Extreme Value Statistics specializes in dealing with extremal events, those of unusual magnitude and rarely observed. It focuses on describing the tail behavior of the underlying distribution governing phenomena, based on the few records falling furthest from the bulk of central data. Results from this framework are ideal tools to extrapolate beyond the sample and allow robust inference on parameters associated to catastrophic events, like exceedance probabilities of high values or return levels over long periods. In my PhD research, I’ve addressed open methodology questions in Extremes, but interesting EV problems can be found everywhere: e.g., in Finance, it might be interesting to ask what is the probability of joint extreme losses for a large number of stocks, leading to a crash of the financial system. Such questions from applied fields have incredible potential to drive multifaceted innovation, rich in both its theoretical and practical dimensions. In this Researcher’s Corner, I invite you to discover more about this field and team up with a Statistician to find an Extreme solution for your Extreme research questions.