Prediction of the critical temperature of a superconductor by using the WOA/MARS, Ridge, Lasso and Elastic-net machine learning techniques

Abstract This study builds a predictive model capable of estimating the critical temperature superconductor from experimentally determined physico-chemical properties material (input variables): features extracted thermal conductivity, atomic radius, valence, electron affinity and mass. original is built using novel hybrid algorithm relied on multivariate adaptive regression splines (MARS) technique in combination with nature-inspired meta-heuristic optimization termed whale (WOA) that mimics social behavior humpback whales. Additionally, Ridge, Lasso Elastic-net models were fitted to same experimental data for comparison purposes. The results current investigation indicate can be successfully predicted this proposed WOA/MARS-based model. Furthermore, obtained are clearly worse than those