Antimicrobial Activity Classification of Imidazolium Derivatives Predicted by Artificial Neural Networks.

The study concludes that the MLP neural network, along with CART and PCA, is a robust tool for predicting the antimicrobial activity class of imidazolium chlorides against Klebsiella pneumoniae. CART and PCA, used in this study, allowed input variable reduction without significant information loss. High classification accuracy and associated metrics affirm the method's potential utility in pre-synthesis assessments, offering valuable insights for antimicrobial compound design.