Model performances
There
were 10 algorithms used to build the ensemble models for A. annuain the research. The AUC and TSS values of each algorithms and ensemble
models were calculated to evaluate their performance under different
climate scenarios.
The TSS results were showed in Table 2. The Average TSS value of RF was
0.835. It got the first place in 10 algorithms, followed by GBM, CTA,
MAXENT, GAM, ANN, MARS, GLM and FDA. The last one was SRE, its average
TSS value was 0.494. The average TSS value for Ensemble was 0.839, it
was greater than any other individual algorithm.
The AUC values were showed in Table3. The average AUC value of RF was
0.964. It was the maximum in 10 algorithms, followed by GBM, MAXENT,
GAM, MARS, CTA, ANN, FDA and GLM. The last one continued to be SRE, its
average AUC value was 0.747. The average AUC value for Ensemble was
0.971, it was greater than any other individual algorithm too.