An Optimized Hamiltonian Quantum Generative Network Based Drug-Drug Interactions Prediction Through Combining Local And Global Features

This work is important in clinical practice and drug development because DDIs are a major factor that poses serious risks to patients’ safety. To enable medication to be given properly with the least side effects, probable DDIs must be predicted correctly. This work introduces a new Spike driven Transformer based Hamiltonian Quantum Generative Adversarial Network with Gorilla Troops Optimizer (SDTbHQGAN-GTO), which achieves every given data set’s accurate DDIs prediction with ease. First, the DDIs datasets are obtained and then pre-processed where noise feature is eliminated from the data through normalization and data cleaning process. Subsequently, the local and global features concerning DDI prediction can be acquired employing Spike driven Transformer (SDT). After that, the DDI interaction can be predicted from the extracted data using Hamiltonian Quantum Generative Adversarial Network with Gorilla Troops Optimizer abbreviated as HQGAN-GTO. It is also possible to optimize the parameters of the HQGAN including reduction of the error and enhancement of the prediction accuracy using the Gorilla Troops Optimizer (GTO). Finally, the performance of the proposed method can be analysed in terms of accuracy, precision, recall and f-score. This approach is superior to the suggested one in terms of the worst-case higher order estimate in offering better 99.82% accuracy, 99.81% precision, 99.812% f1-score, 99.821% recall. Therefore for the accurate prediction of DDIs the proposed method surpasses the current methods in a significant way.