Estimation of parameters in non uniform models on permutations

Sumit Mukherjee · arXiv (Cornell University) · 2013

Abstract. Using large deviation results for a uniformly random permutation in Sn, asymptotics of normalizing constants are computed and limits of permutations obtained for some non uniform distributions on Sn. A pseudo-likelihood type estimator is shown to be consistent in a class of one parameter exponential families on permutations. A new proof of the large deviation principle of a uniformly random permutation on Sn is given. 1.

Read the paper · More papers on PaperTik