ENFA specialization values

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

ENFA specialization values

Aschim, Ruth
Hello,

I am running an ENFA analysis and am having trouble interpreting the specialization values. Do the specialization values correspond to each ecogeographical variable? If so they don't make sense in terms of what I am seeing for the marginality values, i.e. marginality value for water is high and displays a long arrow on the biplot. Therefore, I would think that the specialization value for water would be high, but if the vector of specialization values corresponds to each ecogeographical variable then this is not the case. Am I missing a line of code? An explanation of how to interpret the specialization values from the code and output below would be much appreciated.


slot(maps,"data")[,1:9] <- sqrt(slot(maps,"data")[,1:9])
hist(maps, type = "l")

## Prepare the data for the ENFA
tab <- slot(maps, "data")
pr <- slot(count.points(pipl, maps), "data")[,1]


## Perform the PCA before the ENFA
pc <- dudi.pca(tab, scannf = FALSE)
pc

(enfa1 <- enfa(pc, pr,scannf = FALSE))

enfa1$s #gives specialization values
enfa1$mar#gives marginality values

##OUTPUT##

> (enfa1 <- enfa(pc, pr,scannf = FALSE))
ENFA
$call: enfa(dudi = pc, pr = pr, scannf = FALSE)

marginality: 1.731
eigen values of specialization: 6.61 2.583 1.924 1.49 1.153 ...
$nf: 1 axis of specialization saved

  vector length mode    content
1 $pr    385923 numeric vector of presence
2 $lw    385923 numeric row weights
3 $cw    10     numeric column weights
4 $mar   10     numeric coordinates of the marginality vector
5 $s     9      numeric eigen values of specialization

  data.frame nrow   ncol content
1 $tab       385923 10   modified array
2 $li        385923 2    row coordinates
3 $co        10     2    column coordinates

> enfa1$s  #gives specialization values
[1] 6.6101077 2.5830759 1.9242150 1.4895781 1.1533000 1.0651184 0.8456433 0.7114272 0.4710257

> enfa1$mar #gives marginality values
           Annual         Perennial         Grassland         Deciduous        Coniferous
       0.38442513        0.33468805        0.62931566       -0.31320996       -0.55314134
            Mixed             Water           Wetland             Shrub Distance_to_Roads
      -0.53532160        0.55172030       -0.25751795        0.02533396       -0.11602451

Thank you
~Ruth








        [[alternative HTML version deleted]]

_______________________________________________
R-sig-Geo mailing list
[hidden email]
https://stat.ethz.ch/mailman/listinfo/r-sig-geo
Reply | Threaded
Open this post in threaded view
|

Re: ENFA specialization values

Adrian Dwiputra

Hi Ruth,

 

I had some experience with ENFA 5 years ago. Based on what I wrote back then, I didn’t interpret the specialization values as is, but instead they were aggregated into the overall (global) specialization index and then converted into the tolerance value by simply inverted it (x becomes 1/x). However, the absolute values are the ones that matter in interpreting the specialization coefficient of each EGV: “the higher the absolute value, the more restricted is the range of the focal species on the corresponding variable” Hirzel’s original paper (2002) page 2031 right below the equation (10).

 

Hope this helps a bit.

 

Adrian

 

From: [hidden email]
Sent: Monday, March 23, 2020 8:36 AM
To: [hidden email]
Subject: [R-sig-Geo] ENFA specialization values

 

Hello,

 

I am running an ENFA analysis and am having trouble interpreting the specialization values. Do the specialization values correspond to each ecogeographical variable? If so they don't make sense in terms of what I am seeing for the marginality values, i.e. marginality value for water is high and displays a long arrow on the biplot. Therefore, I would think that the specialization value for water would be high, but if the vector of specialization values corresponds to each ecogeographical variable then this is not the case. Am I missing a line of code? An explanation of how to interpret the specialization values from the code and output below would be much appreciated.

 

 

slot(maps,"data")[,1:9] <- sqrt(slot(maps,"data")[,1:9])

hist(maps, type = "l")

 

## Prepare the data for the ENFA

tab <- slot(maps, "data")

pr <- slot(count.points(pipl, maps), "data")[,1]

 

 

## Perform the PCA before the ENFA

pc <- dudi.pca(tab, scannf = FALSE)

pc

 

(enfa1 <- enfa(pc, pr,scannf = FALSE))

 

enfa1$s #gives specialization values

enfa1$mar#gives marginality values

 

##OUTPUT##

 

> (enfa1 <- enfa(pc, pr,scannf = FALSE))

ENFA

$call: enfa(dudi = pc, pr = pr, scannf = FALSE)

 

marginality: 1.731

eigen values of specialization: 6.61 2.583 1.924 1.49 1.153 ...

$nf: 1 axis of specialization saved

 

  vector length mode    content

1 $pr    385923 numeric vector of presence

2 $lw    385923 numeric row weights

3 $cw    10     numeric column weights

4 $mar   10     numeric coordinates of the marginality vector

5 $s     9      numeric eigen values of specialization

 

  data.frame nrow   ncol content

1 $tab       385923 10   modified array

2 $li        385923 2    row coordinates

3 $co        10     2    column coordinates

 

> enfa1$s  #gives specialization values

[1] 6.6101077 2.5830759 1.9242150 1.4895781 1.1533000 1.0651184 0.8456433 0.7114272 0.4710257

 

> enfa1$mar #gives marginality values

           Annual         Perennial         Grassland         Deciduous        Coniferous

       0.38442513        0.33468805        0.62931566       -0.31320996       -0.55314134

            Mixed             Water           Wetland             Shrub Distance_to_Roads

      -0.53532160        0.55172030       -0.25751795        0.02533396       -0.11602451

 

Thank you

~Ruth

 

 

 

 

 

 

 

 

                [[alternative HTML version deleted]]

 

_______________________________________________

R-sig-Geo mailing list

[hidden email]

https://stat.ethz.ch/mailman/listinfo/r-sig-geo

 


_______________________________________________
R-sig-Geo mailing list
[hidden email]
https://stat.ethz.ch/mailman/listinfo/r-sig-geo