Syracuse Research Computing Article

The Research Computing lab at Syracuse interviewed me about a project I’ve been doing using their High Throughput Computing servers. It’s a case study on applying a Gaussian Process model to factor regression, using R and Stan. As with a Kalman Filter, factor weightings can vary smoothly over time, but using generative modeling in Stan, we are not limited to using updating models with a closed-form solution. The trade-off is that they require a great deal more computing power to fit. Upcoming advances in Stan should speed things up significantly, however.