Quantitative structure-activity relationship based on the interval censored data
Jinsong Zhao · Environmental Chemistry · 2013
In this study,by using Bayesian statistical analysis,quantitative structure-activity relationship(QSAR) models of substituted anilines and phenols for Daphnia magna were developed based on the interval censored data of 24 h acute toxicities.They were compared with QSAR models built from the point estimation,i.e.,the mean and median of the interval censored dataset.The results showed that the former QSAR models could effectively use the biological activity information of specific compounds,and therefore had a better goodness-offit and predictive ability,and a wider range of applications.QSAR model based on the interval censored data could provide a more reliable predicted biological activity for the ecological risk assessment.