<html xmlns:v="urn:schemas-microsoft-com:vml" xmlns:o="urn:schemas-microsoft-com:office:office" xmlns:w="urn:schemas-microsoft-com:office:word" xmlns:x="urn:schemas-microsoft-com:office:excel" xmlns:m="http://schemas.microsoft.com/office/2004/12/omml" xmlns="http://www.w3.org/TR/REC-html40"><head><meta http-equiv=Content-Type content="text/html; charset=us-ascii"><meta name=Generator content="Microsoft Word 12 (filtered medium)"><style><!--
/* Font Definitions */
@font-face
{font-family:Calibri;
panose-1:2 15 5 2 2 2 4 3 2 4;}
@font-face
{font-family:Consolas;
panose-1:2 11 6 9 2 2 4 3 2 4;}
/* Style Definitions */
p.MsoNormal, li.MsoNormal, div.MsoNormal
{margin:0in;
margin-bottom:.0001pt;
font-size:11.0pt;
font-family:"Calibri","sans-serif";}
a:link, span.MsoHyperlink
{mso-style-priority:99;
color:blue;
text-decoration:underline;}
a:visited, span.MsoHyperlinkFollowed
{mso-style-priority:99;
color:purple;
text-decoration:underline;}
p.MsoPlainText, li.MsoPlainText, div.MsoPlainText
{mso-style-priority:99;
mso-style-link:"Plain Text Char";
margin:0in;
margin-bottom:.0001pt;
font-size:10.5pt;
font-family:Consolas;}
span.EmailStyle17
{mso-style-type:personal-compose;
font-family:"Calibri","sans-serif";
color:windowtext;}
span.PlainTextChar
{mso-style-name:"Plain Text Char";
mso-style-priority:99;
mso-style-link:"Plain Text";
font-family:Consolas;}
.MsoChpDefault
{mso-style-type:export-only;}
@page WordSection1
{size:8.5in 11.0in;
margin:1.0in 1.0in 1.0in 1.0in;}
div.WordSection1
{page:WordSection1;}
--></style><!--[if gte mso 9]><xml>
<o:shapedefaults v:ext="edit" spidmax="1026" />
</xml><![endif]--><!--[if gte mso 9]><xml>
<o:shapelayout v:ext="edit">
<o:idmap v:ext="edit" data="1" />
</o:shapelayout></xml><![endif]--></head><body lang=EN-US link=blue vlink=purple><div class=WordSection1><p class=MsoNormal style='margin-bottom:10.0pt;line-height:115%'>Dear colleagues, <o:p></o:p></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>We find the glmm.admb a very interesting tool for ecological modeling. Unfortunately, the at this time still rather sparse documentation of its R-application make for somehow progress in our analysis.<o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'><o:p> </o:p></span></p><p class=MsoNormal style='margin-bottom:10.0pt;line-height:115%'> A few days ago we posted a question regarding the calculation and interpretation of ‘residuals(Model)’ in glmm.admb-R, and Hans Julius Skaug (many thanks Hans) responded. Please see below our initial question; <o:p></o:p></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> > We are using glmmADMB in R to model land cover and water<o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> > quality influence on species diversity of fishes within a study area with<o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> > several subregions.<o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> ><o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> > We defined subregion as a random factor and also ask for individual <o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> > intercepts for the different subregions.<o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> ><o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> > A ‘global’ model for overdispersed count data was formulated as<o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> ><o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> > GM<-glmm.admb( N_Species ~ b1 + b2 + b3 + …+ bn +<o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> Subregion, random = ~ 1,<o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> > group="Subregion", data=input, family="nbinom")<o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> ><o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> > We subsequently evaluated several candidate models that<o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> represent various<o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> > subsets of variables from the global model.<o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> ><o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> > Our input file is A1, with A1$NO representing the observed number of <o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> > species. During the process of examining model performance we used<o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> ><o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> > Observed – Predicted (A1$NO -(fitted(best))) for the<o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> ‘best’ model based<o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> > on lowest AICc to derive residuals for predicted Nspecies. <o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> However, using<o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> > ‘residuals(best)’ produced considerable different (smaller)<o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> values which we<o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> > find somehow puzzling. Are we wrong to assume that<o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> (Nspecies predicted by<o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> > ‘best’ model, + residuals(best)) should add up to Nspecies observed <o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> > (A1$NO)?<o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> ----------------------------------------------------------------------------------------------------------<o:p></o:p></span></p><p class=MsoNormal style='margin-bottom:10.0pt;line-height:115%'>Hans Julius Skaug kindly provided the following answer;<o:p></o:p></p><p class=MsoNormal style='margin-bottom:10.0pt;line-height:115%'>> I think residuals(best) returns<o:p></o:p></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> <o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> [A1$NO -(fitted(best)] / SD<o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> <o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> where SD is the standard deviation, which depends on the distribution <o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> at hand. The code inside glmm.admb that determines SD is:<o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> <o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> tmpsd <- switch(family, poisson = sqrt(lambda), nbinom = <o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> sqrt(lambda *<o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> (1 + lambda/out$alpha)), binomial = sqrt(out$fitted *<o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>> (1 - out$fitted)))<o:p></o:p></span></p><p class=MsoNormal style='margin-bottom:10.0pt;line-height:115%'>> ---------------------------------------------------------------------------------------<o:p></o:p></p><p class=MsoNormal style='margin-bottom:10.0pt;line-height:115%'>Now, since model validation for any application is an essential component of the modeling process, we are asking the ADMB community: would be possible to modify the glmm.admb R-package in the near future so that lambda can be provided in the output? <o:p></o:p></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'><o:p> </o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>Ideally, a desired function to be developed would be a call that provides fitted(model) vs. 'Observed(model)' [=measured data from data input file, something akin to Predicted vs.Observed]. <o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'><o:p> </o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>Also, at current glmm.admb-R output provides alpha as a measure of dispersion of the negative binomial distribution but without a stated lambda value, derivation of SD to calculate a P/O fit is still a challenge to be solved. <o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'><o:p> </o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>I'm sure that other colleagues in ecological research would appreciate such contribution too … <o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'><o:p> </o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>Comments and suggestions are welcome, and thank you for your time. <o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'><o:p> </o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>Cheers, <o:p></o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'><o:p> </o:p></span></p><p class=MsoPlainText><span style='font-family:"Calibri","sans-serif"'>Peter <o:p></o:p></span></p><p class=MsoNormal><o:p> </o:p></p></div></body></html>