SOM Application in Ecological Studies

To demonstrate ecological application of the SOM, the SOM has been applied to a classically simple data set: the distribution of eight tree species at ten sites in southern Wisconsin.

After training the SOM with Wisconsin forest samples, each sample was visualized on the SOM (Figure 8). During the learning process of the SOM, virtual communities are produced. Each virtual community is representative for the sample units assigned to it. Using this virtual community, virtual units (i.e., SOM units) can be classified into corresponding several groups using clustering. To figure out the overall similarities between virtual units, a U-matrix map and a dendrogram of hierarchical cluster analysis is efficient, as shown in Figure 8.

Both methods show similar results on classification of virtual units. There are two main groups: upper areas (cluster I, SUs 1-5) and lower areas (cluster III, SUs 6-10). Each cluster was also subgrouped: cluster I was sub-grouped into cluster Ia (SUs 1 and 2) and cluster Ib (SUs 3-5) and cluster III was subgrouped into cluster IIIa (SUs 9 and 10) and cluster IIIb (SUs 6-8). It is comparable with the dendrogram of cluster analysis with sample units.

Component planes, by visualizing virtual communities on the SOM, are efficient to present importance of each input variables (i.e., species). Figure 9 shows gradient distribution of each species on the SOM. Species bur oak and black oak were abundant in cluster Ia, and white oak and American elm were in cluster Ib. Red oak was abundant in cluster IIIb, whereas ironwood and sugar maple were higher in cluster IIIa. Basswood was abundant in cluster III.

At this point, we have sampling sites and biological variables as parameters on the trained SOM. Using these data, we can summarize the relations between variables. Samples in cluster I are characterized by abundance of bur oak, black oak, white oak, and American elm, whereas samples in cluster III by abundance of red oak, basswood, ironwood, and sugar maple.

Vu3

Vu8

Vu13

II

IIIb

Vu14Vu15 Vu4 Vu5 Vu9 Vu10 Vu3 Vu8 Vu13 Vu1 Vu2 Vu6 Vu7 Vu11Vu12

Figure 8 Classification of virtual units by: (a,b) U-matrix presentation in the SOM and (c) hierarchical clustering analysis.

8 12

Bur oak

Black oak

White oak

Bur oak

3.78

0.50

6.92

5.18

3.44

5.14

2.59

7.06

3.78

0.50

6.92

5.18

3.44

5.14

2.59

0.04

7.98

Black oak

White oak

7.98

Species Distribution Visulization

Figure 9 Visualization of gradient distribution of species in the SOM. The values were calculated during the learning process of the SOM. Dark represents a high value, while light is low.

Figure 9 Visualization of gradient distribution of species in the SOM. The values were calculated during the learning process of the SOM. Dark represents a high value, while light is low.

See also: Artificial Neural Networks: Temporal Networks; Self-Organizing Map.

Trash To Cash

Trash To Cash

This book will surely change your life due to the fact that after reading this book and following through with the steps that are laid out for you in a clear and concise form you will be earning as much as several thousand extra dollars a month,  as you can see by the cover of the book we will be discussing how you can make cash for what is considered trash by many people, these are items that have value to many people that can be sold and help people who need these items most.

Get My Free Ebook


Post a comment