MCMC for Mixture Models: Inferring Earthquake Regimes
The article explores using Markov Chain Monte Carlo (MCMC) methods to infer hidden seismic activity regimes from global earthquake data between 1900 and 2006. It proposes that variations in annual earthquake counts suggest multiple underlying Poisson processes rather than a single random process. A Metropolis-Hastings sampler is implemented to estimate posterior distributions for parameters of a two-component Poisson mixture model, providing uncertainty quantification.
- ▪The dataset includes the number of major earthquakes (magnitude >7) per year from 1900 to 2006, showing high variability inconsistent with a single Poisson process.
- ▪The analysis uses a Bayesian mixture model to identify two potential seismic regimes: a 'quiet' and an 'active' state with different earthquake rates.
- ▪A Metropolis-Hastings MCMC sampler is built from scratch to infer posterior distributions of the model parameters, including Poisson rates and mixing proportions.
Opening excerpt (first ~120 words) tap to expand
try { if(localStorage) { let currentUser = localStorage.getItem('current_user'); if (currentUser) { currentUser = JSON.parse(currentUser); if (currentUser.id === 3843317) { document.getElementById('article-show-container').classList.add('current-user-is-article-author'); } } } } catch (e) { console.error(e); } Berkan Sesen Posted on May 16 • Originally published at sesen.ai MCMC for Mixture Models: Inferring Earthquake Regimes #bayesian #sampling #probabilistic Between 1900 and 2006, the number of major earthquakes per year ranged from 6 to 41. In some decades the planet averaged fewer than 15; in others, closer to 30. That is far too much variation for a single random process. Something changed, and it changed more than once.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at DEV.to (Top).
Discussion
0 commentsMore from DEV.to (Top)
