Identifying diverse metal oxide nanomaterials with lethal effects on embryonic zebrafish using machine learning.

Selection of a single descriptor Random Forest model for external validation The finding that, out of all descriptors used for all previously analysed models, only the Pauling metal atom electronegativity could be reliably considered highly significant prompted investigation of whether an approximate model of excess lethality at 120 hpf could be based purely upon this descriptor.

also in 132,142 other papers