Novel Ensemble Machine Learning Paradigms for the Prediction of Antioxidant Activity of Bryophyllum pinnatum (Lam.) Oken

For herbal and modern drug research, secondary metabolites produced by plants are a valuable source of pharmaceutically active compounds. Northern Nigeria relies heavily on plants as a source of medication. Succulent perennial herb Bryophyllum pinnatum is endemic to Africa and Asia. The current study investigates the in vitro pharmacological potential of the B. pinnatum with the use of predictive data intelligence algorithms to unveil the inhibitory properties of the plant. In vitro biological evaluation of B. pinnatum was carried out. This study explored the utilization of four different data intelligence models composing of two linear models; multi-linear regression (MLR) and step-wise linear regression (SWLR) as well two other nonlinear artificial intelligence (AI)-based models; adaptive neuro-fuzzy inference system (ANFIS) and support vector machine (SVM). Subsequently, four ensemble machine learning approaches including adaptive neuro-fuzzy inference system ensemble (ANFIS-E), support vector machine ensemble (SVM-E), simple average ensemble (SAE) and Gaussian process regression ensemble (GPR-E) were proposed to increase the single model’s accuracy. The ensemble techniques depict an average fitness of 0.999 in both training and testing stages. Moreover, recent and state-of-the-art visualization illustrations were also used for easy grasping and assimilations of the findings. The data-driven approach obtained results showed the AI-based model’s potentiality including SVM and ANFIS in outperforming the classical linear models; SWLR and MLR according to the study's performance metrics. Furthermore, based on the DC assessment criteria investigated in the current work, ensemble machine learning paradigms increased the performance skills of the single algorithms by up to 27% in the testing phase.