Skip to main content
. 2023 Nov 1;14:1283127. doi: 10.3389/fmicb.2023.1283127

Table 3.

Summary statistics of GBM and NN for AFL and FUM.

Overall accuracy test-set Balanced accuracy test-set Overall accuracy validation-set (single year) Balanced accuracy validation-set (single year)
AFL-GBM 96% 96% 96% 50%
AFL-NN 98%* (46/25/15/2)** 98%* 94%* 73%*
FUM-GBM 92% 92% 78% 76%
FUM-NN 96%* (80/65/30/2)** 96%* 73%* 85%*

Overall and balanced accuracy for AFL used 2010 as the independent year and FUM used 2016 (Year was not used as an input variable).

*

Used only input variables with non-zero influence in the GBM.

**

Number of input variables/number of neurons in the first hidden layer/number of neurons in the second hidden layer/number of outputs.