You are here:
A SEMIPARAMETRIC BAYESIAN MODEL FOR CIRCULAR-LINEAR REGRESSION
Citation:
GEORGE, B. J. AND K. GHOSH. A SEMIPARAMETRIC BAYESIAN MODEL FOR CIRCULAR-LINEAR REGRESSION. Communications in Statistics - Simulation and Computation. Taylor & Francis, Inc., Philadelphia, PA, 35(4):911-923, (2006).
Impact/Purpose:
This work presents regression where the goal is to predict a circular response variable from a linear predictor, and it uses a Bayesian framework with a Dirichlet process prior. The illustration relates time of maximum ozone concentration (with time on a 24-hour clock) to the corresponding temperature (the linear predictor). The goal is to predict the time of daily maximum ozone concentration by just knowing the temperature that corresponds to it.
Description:
We present a Bayesian approach to regress a circular variable on a linear predictor. The regression coefficients are assumed to have a nonparametric distribution with a Dirichlet process prior. The semiparametric Bayesian approach gives added flexibility to the model and is useful especially when the likelihood surface is ill-behaved. Markov chain Monte Carlo techniques are used to fit the proposed model and generate predictions. The method is illustrated using an environmental data set.