是否评估了筛选中使用的预测的质量？

是否评估了筛选中使用的预测的质量？

Is the quality of the predictions used in screening assessed?

算法会在适当时考虑适用领域。此外，通常会组合使用多个模型，以检查其一致性。对于预期误差可能较大的模型，预测仅用作其他发现的支持信息，例如注册档案中报告的实验研究中存在的微弱/不确定的危害证据。所有发现（包括模型预测）的有效性和相关性均由成员国在人工筛选过程中进行评估。

The algorithms use applicability domain considerations when appropriate. Moreover, multiple models are often used in combination to be able to check their consistency. For models where the expected error may be larger, predictions are only used as supporting information to other findings, such as weak/inconclusive evidence of hazard in experimental studies reported in the registration dossiers. The validity and relevance of all findings, including the model predictions are assessed by Member States during manual screening.