             SOME BAYESIAN INFERENCES FOR VONMISES DISTRIBUTION K. MURALIDHARAN 1 and RAJIV PARIKH 2 1 Department of Statistics, Faculty of Science, The M. S. University of Baroda, Vadodara- 390 002, India. Email: lmv_murali@yahoo.com 2 Department of Statistics, Bhavnagar University, Bhavnagar- 364 002, India. ABSTRACT: The von Mises distribution (VMS) is considered to be a circular distribution having two parameters and is the natural analogue on the circle of the Normal distribution on the real line. Since the density function involves the modified Bessel function, quite often the sample generation becomes tedious. This limits the applicability of the distribution. In this paper, we have done some Bayesian inferences of the parameters using weighted bootstrap resampling method and a modified version of Gibb’s sampling under different prior distributions. A numerical example is also presented. Keywords: markov chains; resamplig techniques; prior distribution; posterior distribution; weighted bootstrap; mean direction; concentration parameter. 1. INTRODUCTION Von Mises distribution is a symmetric unimodal distribution, which is the most common model for unimodal samples of circular data. It is denoted by VM(μ,κ) and its probability density function is given by