Talk:Zero-inflated model

Insurance example is misleading
The insurance example decreases the value of this article as it is misleading. Insurance companies have a sustainable business model because the rate of claims is low, not because they zero-inflate the distribution. A valid (but still bad) example would be an insurance company that has a clause buried deeply in their T&Cs that prevents a random set of claims from being paid out. 2.31.1.97 (talk) 21:06, 2 May 2014 (UTC)
 * I reworded the example. I think it is now correct, although perhaps still not the ideal example. Rlendog (talk) 21:53, 2 May 2014 (UTC)
 * Thanks. While removing the bankruptcy claim is an improvement, my main criticism is that this example does nothing to illustrate the problem and is indeed misleading. I've changed the example to hopefully provide better insight. 2.31.1.244 (talk) 13:41, 18 May 2014 (UTC)

Dr. Winkelmann's comment on this article
Dr. Winkelmann has reviewed this Wikipedia page, and provided us with the following comments to improve its quality:

"Unclear meaning: "concerns a random event containing excess zero-count data in unit time"

better: "for count data with excess zeros".

First formula: Unclear what the subcript "j" is for. Use "i" throughout.

No-one in practice uses the two-parameter ZIP model. It almost always involves covariates/regressors. Model and estimation part should reflect that (Maximum Likelihood).

Sentence starting with "If the count data" is grammatically incorrect.

The subsection titled "Discrete pseudo compound Poisson model" should be dropped. This has no relation to the topic of the entry "Zero-inflated model". Zero-inflated models are mixture models and not compound models. The paper this subsection relies on describes a class of compound models also known as  stopped-sum distributions (see e.g. Santos Silva and Windmeijer). This belongs elsewhere.

Finally: the entry continues the mistaken tradition of attributing the zero-inflated Poisson model to Lambert (1992), when it was in fact introduced  into the literature by Mullahy (1986), as a regression model, or even earlier by  Johnson and Kotz as a pure statistical distribution.

Santos Silva, J. M.C. and Windmeijer, Frank, Two-Part Multiple Spell Models for Health Care Demand. Journal of Econometrics, Vol. 104, August 2001. Available at SSRN: http://ssrn.com/abstract=276962

Mullahy, J. (1986). Specification and testing in some modified count data models, Journal of Econometrics, 33, 341-365."

We hope Wikipedians on this talk page can take advantage of these comments and improve the quality of the article accordingly.

We believe Dr. Winkelmann has expertise on the topic of this article, since he has published relevant scholarly research:


 * Reference : Gregori Baetschmann & Rainer Winkelmann, 2014. "A dynamic hurdle model for zero-inflated count data: with an application to health care utilization," ECON - Working Papers 151, Department of Economics - University of Zurich.

ExpertIdeasBot (talk) 16:56, 27 July 2016 (UTC)