A novel hybrid Harris hawks optimization and support vector machines for drug design and discovery

Cheminformatics has main research factors due to increasing size of the search space of chemical compound databases and the importance of similarity measurements for drug design and discovery. Several traditional methods are used to predict drug design and discovery, which are relatively less efficient and weak. This paper proposes two classification approaches called HHO-SVM and HHO-kNN, which hybridizes a novel metaheuristic algorithm called Harris hawks Optimization (HHO) with Support Vector Machines (SVM) and the k-Nearest Neighbors (k-NN) for chemical descriptor selection and chemical compound activities. The core exploratory and exploitative processes of HHO is adapted to select the significant features for achieving high classification accuracy. Two chemical datasets (MonoAmine Oxidase (MAO) and QSAR Biodegradation) are used in the experiments. The experimental results proved that the proposed HHO-SVM approach achieved the highest capability to obtain the optimal features compared with several well-established metaheuristic algorithms including: Particle Swarm Optimization (PSO), Simulated Annealing (SA), Dragonfly Algorithm (DA), Butterfly Optimization Algorithm (BOA), Moth-Flame Optimization Algorithm (MFO), Grey Wolf Optimizer (GWO), Sine Cosine Algorithm (SCA), and Slap Swarm Algorithm (SSA).