Features |
Features |
A |
B |
C |
D |
E |
F |
G |
H |
I |
J |
K |
L |
Data Inputs |
Abundance records can be employed as training
data |
|
|
|
|
|
ü |
|
ü |
|
|
ü |
ü |
|
Flexibility to accommodate pools of biotic and abiotic variables as
training data |
|
|
|
|
|
ü |
|
ü |
|
|
ü |
|
|
Categorical and continuous variables can be included as training data |
|
|
|
|
|
ü |
|
ü |
|
|
ü |
|
|
Categorical and or continuous predictor data can be employed |
ü |
|
|
|
|
ü |
|
|
ü |
ü |
|
|
|
Large datasets efficiently handled |
ü |
ü |
|
|
|
|
ü |
|
|
|
|
|
|
Traits, phylogeny, and other ancillary data can be incorporated |
|
|
|
|
|
|
|
ü |
|
|
|
|
Implementation |
Model outputs easy to visualize, understand,
and interpret |
ü |
ü |
|
|
|
|
|
|
|
|
|
|
|
Readily available tutorials and examples of code |
ü |
ü |
ü |
ü |
ü |
|
|
ü |
ü |
|
|
|
|
Technically challenging to implement and fit models |
|
|
|
|
|
ü |
ü |
ü |
|
|
|
|
|
Computationally demanding and time consuming to run |
|
|
ü |
ü |
ü |
ü |
|
ü |
|
|
|
|
|
Results can be challenging to interpret |
|
|
ü |
ü |
ü |
ü |
ü |
ü |
|
ü |
|
|
|
Inscrutable or “black box” approach |
|
|
|
|
|
ü |
ü |
|
|
|
|
|
Analytical Properties |
Allows for non-linear responses |
ü |
ü |
|
ü |
ü |
ü |
ü |
ü |
ü |
ü |
ü |
ü |
|
Allows for interaction between predictor variables |
ü |
|
ü |
ü |
ü |
ü |
ü |
ü |
|
|
ü |
|
|
Accounts for unexplained co-occurrences |
|
|
|
|
|
|
ü |
ü |
|
|
|
ü |
|
Uncertainty can be explicitly quantified |
|
|
|
|
|
|
|
ü |
|
|
ü |
ü |
|
Predicts compositional differences between locations as a continuous
function of environmental gradients |
|
|
|
|
|
|
|
|
ü |
ü |
|
|
|
Dissimilarities between pairs of sites related to differences in
environmental conditions and geographic (distance) isolation |
|
|
|
|
|
|
|
|
ü |
ü |
|
|
Performance |
Moderate to high predictive power |
|
|
|
|
|
|
|
ü |
ü |
|
|
|