Abstract
A quantitative structure–bioconcentration factor (BCF) relationship was developed to predict the BCF of some polychlorinated biphenyls (PCBs). A set of 1,497 zero- to three-dimensional descriptors were used for each molecule in the data set. Multivariate adaptive spline was successfully used as a descriptor selection method. Two descriptors were selected and used as inputs for adaptive neuro-fuzzy inference system (ANFIS). The root mean square errors for the calibration, prediction, and validation sets were 0.16, 0.24, and 0.23, respectively. The results were compared with those obtained from two other models. In one model, multivariate adaptive regression spline (MARS) was used as a descriptor selection method and also as a mapping model. In the other model, after selection of descriptors by MARS, multiple linear regression (MLR) was applied for modeling. The results showed MARS–ANFIS can be used as a powerful model for prediction of BCFs of PCBs.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call