[R-sig-eco] Ranked abundance distribution

[Martin Weiser]

Sol Noetinger píše v Út 17. 12. 2013 v 12:01 -0300:
> Hello,
> 
> I am trying to apply different statistics methods in a field that traditionally is not very keen to it and in consequence I am trying to learn all that I can. 
> To the point, I am studying a palynological succession from the Devonian. I have counts of palynomorphs (around 250) from a set of 17 samples. I use the relative abundance to standardise the counting since there are some samples that did have not enough specimens.
> I have tested cluster analysis with different packages, resulting in two clear groups. I tested the abundance distribution on both groups to see which model fits better.
> 
> This is the summary:
> 
> Cluster I
> RAD models, family poisson 
> No. of species 24, total abundance 100
> 
>            par1     par2    par3   Deviance AIC     BIC    
> Null                               55.2189      Inf     Inf
> Preemption    0.1                  85.4721      Inf     Inf
> Lognormal   0.20534  1.6811         8.0522      Inf     Inf
> Zipf        0.42497 -1.4264         1.4461      Inf     Inf
> Mandelbrot  1.4285  -1.8885      1  3.4265      Inf     Inf
> 
> Cluster II
> RAD models, family poisson 
> No. of species 35, total abundance 100
> 
>            par1     par2    par3   Deviance AIC     BIC    
> Null                               25.7004      Inf     Inf
> Preemption    0.1                  27.8760      Inf     Inf
> Lognormal   0.21756  1.3473         4.7797      Inf     Inf
> Zipf        0.27724 -1.0959         4.9038      Inf     Inf
> Mandelbrot  0.64175 -1.3825      1  4.9181      Inf     Inf
> 
> I read from the manual that to see which models fits better you use the AIC values. 
> What is the meaning of getting "infinite"?
> Can I use the Deviance value to compare the models?
> And in case I can use the deviance, since there are very close values, should I run a test to see if the differences are significant? in that case, which  one?.
> 
> I apologise if my questions are too basic, or if I should refer to a different kind of forum or thread. 
> I hope you can help me, thank you for your time,
> Regards,
> 
> Sol
> 
> 
> 	[[alternative HTML version deleted]]
> 
> _______________________________________________
> R-sig-ecology mailing list
> R-sig-ecology at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-ecology

Hi Sol,

I am not sure whether I got it right:
Your data are exotic, but besides of this, you just need simple
clustering of 17 samples with cca. 250 "species".

If this is a case, what about to use e.g. vegdist from the vegan package
to get inter-sample distances and then run hclust?
Then the problem reduces to finding appropriate distance measure - if
you have no better idea, one may start with something like Bray-Curtis 

But maybe I am wrong and somebody will correct me.
HTH,
Martin