Software
- Sample data generation using Chinese Restaurant Process and posterior inference using collapsed Gibbs sampler for Dirichlet Process mixture of multinomials, based on "Modeling individual differences using Dirichlet processes", Navarro, Griffiths, Steyvers, Lee, 2005. Linux C++ source; requires Google gflags and the GNU Scientific Library.
- R Code to accompany Hierarchical Dirichlet Processes project in topic modeling.
- Markov Chain Monte Carlo simulation engine for modeling data as coming from a Student's t distrubution. Uses MCMC to draw samples from a posterior distribution of the parameters - namely the degrees of freedom nu, the location mu, and the scale sigma - given an input data set Y. Joint work with Professor Draper, UC Santa Cruz. Linux C++ source; requires Google gflags and the GNU Scientific Library.