Material Detail

A Flexible Model for Count Data: The COM-Poisson Distribution

A Flexible Model for Count Data: The COM-Poisson Distribution

This video was recorded at Solomon seminar. Count data arise in many contexts, from word lengths to traffic volume to number of bids in online auctions, and generally in many event-counting applications. Yet, there is a scarcity of statistical models for such data. The Poisson distribution is the most popular distribution for modeling count data, yet it is constrained by its equi-dispersion assumption, making it less than ideal for modeling real data that often exhibit over-dispersion or under-dispersion. The COM-Poisson distribution is a two-parameter generalization of the Poisson distribution that allows for a wide range of over-dispersion and under-dispersion. It also contains the Bernoulli and geometric distributions as special cases, and as a member of the exponential family has useful statistical properties. This distribution's flexibility and special properties have prompted a fast growth of methodological and applied research in various fields. In this talk, I will introduce the COM-Poisson distribution and regression model and mention several other COM-Poisson models that have been published thus far. I will also describe applications of the COM-Poisson in various areas including disclosure limitation, marketing, transportation and linguistics.

Quality

  • User Rating
  • Comments
  • Learning Exercises
  • Bookmark Collections
  • Course ePortfolios
  • Accessibility Info

More about this material

Comments

Log in to participate in the discussions or sign up if you are not already a MERLOT member.