[R] Separation issue in binary response models - glm, brglm, logistf
Xochitl.Cormon at ifremer.fr
Thu Feb 28 17:56:59 CET 2013
Le 28/02/2013 17:22, Ben Bolker a écrit :
Thank you for your help !
> Xochitl CORMON<Xochitl.Cormon<at> ifremer.fr> writes:
>> Dear all,
>> I am encountering some issues with my data and need some help. I am
>> trying to run glm analysis with a presence/absence variable as
>> response variable and several explanatory variable (time,
>> location, presence/absence data, abundance data).
>> First I tried to use the glm() function, however I was having 2
>> warnings concerning glm.fit () : # 1: glm.fit: algorithm did not
>> converge # 2: glm.fit: fitted probabilities numerically 0 or 1
>> occurred After some investigation I found out that the problem was
>> most probably quasi complete separation and therefor decide to use
>> brglm and/or
>> * logistf : analysis does not run When running logistf() I get a
>> error message saying : # error in chol.default(x) : # leading minor
>> 39 is not positive definite I looked into logistf package manual,
>> on Internet, in the theoretical and technical paper of Heinze and
>> Ploner and cannot find where this function is used and if the error
>> can be fixed by some settings.
> chol.default is a function for Cholesky decomposition, which is going
> to be embedded fairly deeply in the code ...
If I understand good I should just not use this package as this error is
not easily fixable ?
>> * brglm : analysis run However I get a warning message saying : #
>> In fit.proc(x = X, y = Y, weights = weights, start = start,
>> etastart # = etastart, : # Iteration limit reached Like before i
>> cannot find where and why this function is used while running the
>> package and if it can be fixed by adjusting some settings.
>> In a more general way, I was wondering what are the fundamental
>> differences of these packages.
> You might also take a crack with bayesglm() in the arm package, which
> should (?) be able to overcome the separation problem by specifying a
> not-completely-uninformative prior.
Thank you for the tip I will have a look into this package and its doc
tomorrow. Do you have any idea of what is this fit.proc function ?
>> I hope this make sense enough and I am sorry if this is kind of
>> statistical evidence that I'm not aware of.
Here an extract of my table and the different formula I run :
>>> head (CPUE_table)
>> Year Quarter Subarea Latitude Longitude Presence.S CPUE.S
>> Presence.H CPUE.H Presence.NP CPUE.NP Presence.BW CPUE.BW
>> Presence.C CPUE.C Presence.P CPUE.P Presence.W CPUE.W 1 2000 1 31F1
>> 51.25 1.5 0 0 0 0 0 0 0 0 1 76.002 0 0 1 3358.667
>> logistf_binomPres<- logistf (Presence.S ~ (Presence.BW + Presence.W
>> + Presence.C + Presence.NP +Presence.P + Presence.H +CPUE.BW +
>> CPUE.H + CPUE.P + CPUE.NP + CPUE.W + CPUE.C + Year + Quarter +
>> Latitude + Longitude)^2, data = CPUE_table)
>> Brglm_binomPres<- brglm (Presence.S ~ (Presence.BW + Presence.W +
>> Presence.C + Presence.NP +Presence.P + Presence.H +CPUE.BW + CPUE.H
>> + CPUE.P + CPUE.NP + CPUE.W + CPUE.C + Year + Quarter + Latitude +
>> Longitude)^2, family = binomial, data = CPUE_table)
> It's not much to go on, but:
Yeah sorry my table header appeared really bad on the email :s
> * are you overfitting your data? That is, do you have at least 20
> times as many 1's or 0's (whichever is rarer) as the number of
> parameters you are trying to estimated?
I have 16 explanatory variable and with interactions we go to 136
> length (which((CPUE_table)[,]== 0))
> length (which((CPUE_table)[,]== 1))
I assume the over fitting is good, isn't it?
> * have you examined your data graphically and looked for any strong
> outliers that might be throwing off the fit?
I did check my data graphically in a lot and different ways. However if
you have any particular suggestions, please let me know. Concerning
strong outliers, I do not really understand what you mean. I have
outliers here and there but how can I know that they are strong enough
to throw off the fit? Most of the time they are really high abundance
coming from the fact that I'm using survey data and probably related to
the fact that the boat fished over a fish school.
> * do you have some strongly correlated/multicollinear predictors?
It's survey data so they indeed are correlated in time and space.
However I checked the scatterplot matrix and I didn't notice any linear
relation between variable.
> * for what it's worth it looks like a variety of your variables
> might be dummy variables, which you can often express more compactly
> by using a factor variable and letting R construct the design matrix
> (i.e. generating the dummy variables on the fly), although that
> shouldn't change your results
I will check about dummy variable concept as to be honest I don't really
understand what it means...
Thank you again for your time and help
> ______________________________________________ R-help at r-project.org
> mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do
> read the posting guide
> and provide commented, minimal, self-contained, reproducible code.
More information about the R-help