Do you mean rare event logistic regression? If so, you could try the Zelig package (developed by Gary King and others), which allows you to do such analyses.

Thinking of rare events, if what you are looking to do is regressions, you might simply need to do a generalised linear model regression that uses the "beta-binomial" distribution model, instead of regular binomial... Seems it is good for situation where the number of "successes" (events, 1's) is either very small (rare) or unusually high? There must be several packages in R that can switch to this type of distribution.

Actually, I want SKEW LOGISTIC DISTRIBUTION of Wahed and Ali (2001) [15. Wahed, A.S., Ali, M. Masoom., 2001. The skew logistic distribution, J. Stat. Res., 35, 71–80.]

## All Answers (6)

Rense Nieuwenhuis· Stockholm UniversityJoseph Tkacz· Health Analyticshttp://gking.harvard.edu/files/gking/files/0s.pdf

Ujjwal Das· Indian Institute of Management UdaipurDaniel Gallant· Parks CanadaPartha Jyoti Hazarika· Dibrugarh UniversityYuxin Chen· University of ZurichCan you help by adding an answer?