Nonparametric multimodal regression for circular data

Multimodal regression estimation methods are introduced for regression models involving circular response and/or covariate. The regression estimators are based on the maximization of the conditional densities of the response variable over the covariate. Conditional versions of the mean shift and the circular mean shift algorithms are used to obtain the regression estimators. The asymptotic properties of the estimators are studied and the problem of bandwidth selection is discussed.