194
Fig. 3 Average overall accuracy with confidence limits based on tenfold cross-validation with five
repetitions. The stars denote the overall accuracy derived by a null-model. Stars within confidence
limits signify models that were not significantly better than the null-model and thus do not provide
credible results. Note the very high range of confidence limits for the infrequent species data set
(Spinfreq)
Fig. 4 Variable importance of the ten most important image parameters for the three species subsets, i.e. all species (a), frequent species (b) and infrequent species (c). For full names of image
parameters see Table 2
J. Oldeland et al.
Fig. 3 Average overall accuracy with confidence limits based on tenfold cross-validation with five
repetitions. The stars denote the overall accuracy derived by a null-model. Stars within confidence
limits signify models that were not significantly better than the null-model and thus do not provide
credible results. Note the very high range of confidence limits for the infrequent species data set
(Spinfreq)
Fig. 4 Variable importance of the ten most important image parameters for the three species subsets, i.e. all species (a), frequent species (b) and infrequent species (c). For full names of image
parameters see Table 2
J. Oldeland et al.