Effects of correcting for geographical sampling bias on the predictive performance of New Zealand tree fern distribution models trained on herbarium and NVS datasets.

MaxEnt was used to fit the models (feature type = LQ) and model performance indicated by mean (±1 standard deviation) AUC and COR values, evaluated by using the independent LUCAS dataset.采用最大熵模型（MaxEnt）拟合模型，特征类型设置为LQ；模型性能以均值（±1标准差）的受试者工作特征曲线下面积（Area Under the Receiver Operating Characteristic Curve，AUC）与相关系数（Correlation，COR）值表征，所有性能指标均通过独立LUCAS数据集完成评估。