GLIM (software)

GLIM (an acronym for Generalized Linear Interactive Modelling) is a statistical software program for fitting generalized linear models (GLMs).

It was developed by the Royal Statistical Society's

Working Party on Statistical Computing

(later renamed the GLIM Working Party),{{cite web |url=http://www.rss.org.uk/main.asp?page=2128 |title=Royal Statistical Society webpage on Working Parties |accessdate=2007-12-18 |url-status=bot: unknown |archiveurl=https://web.archive.org/web/20070221234751/http://www.rss.org.uk/main.asp?page=2128 |archivedate=February 21, 2007 }}

chaired initially by John Nelder.

{{cite journal |last=Nelder |first=John |authorlink=John Nelder |year=1975 |title=Announcement by the Working Party on Statistical Computing: GLIM (Generalized Linear Interactive Modelling Program) |journal=Journal of the Royal Statistical Society, Series C |volume=24 |issue=2 |pages=259–261 |jstor= 2346575 }}

It was first released in 1974

with the last major release, GLIM4, in 1993.

{{cite book |title=The GLIM System: Release 4 Manual |first=Brian | last=Francis |author2=Mick Green |author3=Clive Payne |year=1993 |publisher=Clarendon Press |location=Oxford |isbn= 0-19-852231-2}}

GLIM was distributed by the Numerical Algorithms Group (NAG).{{cite web |url=http://www.nag.co.uk/stats/GDGE_soft.asp |title=Generalized Linear Interactive Modeling Package (GLIM) |accessdate=2007-12-18 |url-status=bot: unknown |archiveurl=https://web.archive.org/web/20101012041544/http://www.nag.co.uk/stats/GDGE_soft.asp |archivedate=12 October 2010 }}

GLIM was notable for being the first package capable of fitting a wide range of generalized linear models in a unified framework, and for encouraging an interactive, iterative approach to statistical modelling.

{{cite book | title=Statistical Modelling in GLIM | first=Murray | last=Aitkin | author2=Dorothy Anderson | author3=Brian Francis | author4=John Hinde | year=1989 | publisher=Oxford University Press | location=Oxford | isbn=0-19-852203-7 | url=https://archive.org/details/statisticalmodel00oces }}

GLIM used a command-line interface and allowed users to define their own macros. Many articles in academic journals were written about the use of GLIM.{{cite journal|last=Wacholder|first=Sholom|title=Binomial regression in GLIM: Estimating risk ratios and risk differences|journal=American Journal of Epidemiology|year=1986|volume=123|issue=1|pages=174–184 |pmid=3509965}}{{cite journal|last=Aitken|first=Murray|author2=Clayton, David |authorlink2=David Clayton |title=The Fitting of Exponential, Weibull and Extreme Value Distributions to Complex Censored Survival Data Using GLIM|journal=Journal of the Royal Statistical Society, Series C|year=1980|volume=29|issue=2|pages=156–163|jstor=2986301}}{{cite journal|last=Aitkin|first=Murray|title=Modelling Variance Heterogeneity in Normal Regression Using GLIM|journal=Journal of the Royal Statistical Society, Series C|year=1987|volume=36|issue=3|jstor=2347792}}{{cite journal|last=Whitehead|first=John|title=Fitting Cox's Regression Model to Survival Data using GLIM|journal=Journal of the Royal Statistical Society, Series C |year=1980|volume=29|issue=3|jstor=2346901}}{{cite journal|last=Berman, Mark |author2=Turner, Rolf T.|title=Approximating Point Process Likelihoods with GLIM|journal=Journal of the Royal Statistical Society, Series C|year=1992|volume=41|issue=1|pages=31–38|jstor=2347614}}{{cite journal|last=Decarli, A.|author2=La Vecchia, C.|title=Age, period and cohort models: review of knowledge and implementation in GLIM|journal=Rev. Stat. App.|year=1987|volume=20|pages=397–409}}{{cite journal|doi=10.2307/1403047|last=Jørgensen|first=Bent|title=The Delta Algorithm and GLIM |journal=International Statistical Review / Revue Internationale de Statistique|year=1984|volume=52|issue=3|pages=283–300|jstor=1403047}} Two GLIM conferences were held in London (1982) and Lancaster (1985) and the Statistical Modelling Society, with its annual workshops, grew out of them. GLIM was reviewed in The American Statistician in 1994, along with other software for fitting generalized linear models.{{cite journal|doi=10.2307/2684732|last=Hilbe|first=Joseph|authorlink=Joseph Hilbe|title=Review: Generalized Linear Models |journal=The American Statistician|year=1994|volume=48|issue=3|pages=255–265|jstor=2684732|arxiv=1308.2408}}

The GLIMPSE system was later developed to provide a knowledge based front-end for GLIM.{{Cite journal| volume = 1| journal = Knowledge-Based Systems| pages = 173| year = 1988| doi = 10.1016/0950-7051(88)90075-5| last1 = Wolstenholme| title = GLIMPSE: a knowledge-based front end for statistical analysis | first1 = D.| last2 = Obrien | first2 = C. | first3 = J.| last3 = Nelder| issue = 3}}

GLIM is no longer actively developed or distributed.