paul-buerkner/brms

Conway-Maxwell-Poisson distribution

Open

#607 建立於 2019年2月22日

在 GitHub 查看
 (34 留言) (5 反應) (0 負責人)R (1,402 star) (220 fork)batch import
familyfeaturegood first issue

描述

The Conway-Maxwell-Poisson distribution (https://en.wikipedia.org/wiki/Conway%E2%80%93Maxwell%E2%80%93Poisson_distribution) is a generalization of the Poisson distribution that allows for both over- and underdispersion. There have been some attempts to get this distribution running in Stan, but I haven't seen one that works fast enough for most practical applications. Once we got such an implementation though, I think the distribution would make a very valuable addition to brms.

貢獻者指南

技術棧
無
領域
datamachine learning
議題類型
feature
難度面向新貢獻者的預計實作難度，1 表示很小改動，5 表示專家級工作。
5
預計時間有經驗貢獻者完成調查、實作、測試並準備 pull request 的粗略時間範圍。
over 1 week
活動狀態議題目前的可參與程度：新鮮、活躍、陳舊、阻塞或等待維護者輸入。
stale
清晰度議題是否清楚說明預期改動、驗收標準和下一步。
needs investigation
前置要求
Bayesian statisticsStan modelingR programmingProbability distributions
新手友善度1-100 的估計分數，表示該議題對首次貢獻者的友善程度。
15
研究方向
Investigate the Conway Maxwell Poisson distribution (see Wikipedia link in issue). Review previous attempts to implement it in Stan and identify why they were too slow. Focus on developing an efficient sampling or likelihood computation algorithm in Stan. Once a fast implementation is available, integrate it into brms by adding a new family function, following existing patterns in the package. Test on simulated data and compare to standard Poisson.