A survival mixture cure model in smartphone-based earthquake early warning
Francesco Finazzi (University of Bergamo)
Wednesday 8th November 13:00-14:00 Maths 311B
Crowdsourced smartphone-based earthquake early warning systems have recently emerged as reliable alternatives to more expensive solutions based on scientific instruments. During the deadly Turkish-Syrian event of 6 February 2023, the system implemented by the Earthquake Network citizen science initiative provided up to 58 seconds of warning to people exposed to life-threatening ground shaking. This seminar will present a statistical methodology based on a survival mixture cure model that allows full Bayesian inference of earthquake epicentre, depth and time of origin using data collected by the smartphones. Model estimation is based on an efficient tempering MCMC algorithm to deal with the multi-modality of the posterior distribution.