Guide to obtain AIC (Akaike Information Criterion) with NicheA

Guide to obtain AIC (Akaike Information Criterion) with NicheA

Nichea 3.0 implemented an option to the “model selection” tool in the ENMTOOLS software that can be employed for several model algorithms (e.g., GLM, NicheA, Maxent).

In NicheA, select “Calculate AIC/BIC value” from the Toolbox menu.

Figure 1. The menu to invoke this function

A window will appear with some options. Here we have to choose the proper files to calculate de AIC index. We will need to fill the textbox’s at this order: occurrence points in.csv extension, the Maxent output in .asc file, and then the .lambdas file from the Maxent model.

Figure 2. The dialog for this function

Figure 3. Detail of the dialog for this function.

  • (1) The occurrence points in CSV format with the same format used in the other functions in NicheA, compatible with Maxent format.
  • (2) Using the “Add” button we select the .asc or .tif file obtained after the Maxent analysis in raw format. For models other than Maxent, raster values must be rescale to sum 1 for the entire raster. You can use "Change the range of ENM's result" function to finish this step.
  • (3) Use the “Get from Maxent’s ‘.lambdas’ file” to select the .lambdas file from the Maxent analysis.
    The “number of parameters” used by this file will appear, in this case “2.” You can explore the number o Maxent’s parameters by opening the .lambdas file with a text editor and look for the parameters used. For models other than Maxent you can write the number of parameters in this section.

    • Figure 4. A sample .lambdas file generated in Maxent.

    Finally we select “OK” and then we wait for the pop up window with the result:

    Figure 5. The result for this function