0

I have n items ranked by k experts. I would like to aggregate the ranks using a Bayesian model. However, I do not know if there is a way to model the ranks.

P.S.: You can assume that we have a feature vector for each item, although I prefer that there is no such an assumption.

Majid Mohammadi
  • 99
  • Can you give more information? Which technique do you want to use to estimate the Bayesian model? Bayesian models work with distributions, but I don't see any distributions in your question.. –  Nov 01 '18 at 10:20
  • Thanks for your reply. I would like to develop a model myself. What I need to first model the ranking data using a distribution. – Majid Mohammadi Nov 01 '18 at 15:21
  • For Bayesian Statistics you need a likelihood which you can obtain by using MLE on the data you've got (the ranks), you also need priors (expectations for the outcome) and than you can estimate a posterior, which is the model you want to develop I think. All those things: likelihood, priors and posterior are probability distributions. –  Nov 01 '18 at 15:25

0 Answers0