Distributed heterogeneous ensemble learning on Apache Spark for ligand-based virtual screening

Virtual screening is one of the most common computer-aided drug design techniques that apply computational tools and methods on large libraries molecules to extract drugs. Ensemble learning a recent paradigm launched improve machine results in terms predictive performance robustness. It has been successfully applied ligand-based virtual (LBVS) approaches. Applying ensemble huge molecular computationally expensive. Hence, distribution parallelisation task have become significant step by using sophisticated frameworks such as Apache Spark. In this paper, we propose new approach HEnsL_DLBVS, for heterogeneous learning, distributed Spark large-scale LBVS results. To handle problem imbalanced big training datasets, novel hybrid technique. We generate datasets evaluate approach. Experimental confirm effectiveness our with satisfactory accuracy its superiority over homogeneous models.