Artificial neural network-based QSAR model for predicting degradation techniques of pharmaceutical contaminants in water bodies with experimental verification

Abstract

The pharmaceutical industry has been increasing its production, manufacturing, and promotion of various products, resulting in a rise in contaminants in water. Drugs pose a significant threat because they can persist in water for extended periods. To address this issue, a project was initiated to develop a model for predicting the degradation percentage of pharmaceutical contaminants in water using different physicochemical methods. The model is based on artificial neural networks and uses quantitative structure–activity relationship (QSAR) to predict the degradation percentage of drugs in water when subjected to ozonation, ozonation + H₂O₂, activated carbon use, UV radiation, Fenton darkness, and photo-Fenton + H₂O₂. A total of 75 models were developed, and five met the validation criteria. With the help of the validated models, the study predicted the elimination percentages of more prevalent drugs in water sources. The results reveal that ozonation, with or without peroxide, is the best degradation method. The study has successfully verified the predicted results by conducting experiments on the degradation of an aqueous solution of cephalexin using ozonolysis, which resulted in a degradation percentage of 97.8%. The industry can use the ANN-INQA algorithm to select an optimal method for effectively degrading pharmaceutical contaminants, which can help reduce costs and save time.