Abstract by Joshua Fullwood
Statistical Approximation of Uncertain Data for Random Walk Bayesian Monte Carlo
Solving inverse problems for 19th century seismic events requires abstraction of data that was not observed directly. Additionally, many types of data necessary for modern geophysical modeling requires measurements that are uncertain by nature. We discuss the use of kernel density estimators (KDE) in the construction of a prior distribution from modern USGS data. We also discuss different kinds of sensitivity analyses to benchmark the effect of different approximations on the results of our model.