Application of machine learning models to predict ecotoxicity of ionic liquids (Vibrio fischeri) using VolSurf principal properties

Abstract. Owing to the rapid growth in IL synthesis due to feasible cation–anion combinations, knowledge of their toxicity is pertinent for their successful application. Toxicity information measurement of various ILs on a broad spectrum of conditions through experimental techniques is way demanding on time, resources, and is at times impractical. Various research works have been performed in Quantitative Structure Activity/Property Relationship (QSAR/QSPR) for IL toxicity prediction. In this study, ML models have been trained and tested on Vibrio fischeri toxicity data set using in silico principal properties (PPs) as descriptors. Deploying this properties aid in considering both the effect of cations and anions on Vibrio fischeri toxicity prediction. Among the models trained, the Random Forest model proved to be the most precise nevertheless, decision tree model was the most accurate and consistent. Considering the importance of the descriptors to Vibrio fischeri toxicity selection techniques and model optimization.

UTP CRIS Expert