Leveraging Machine Learning to Predict High-Temperature Superconductors

High-temperature superconductors hold transformative potential across quantum computing, power transmission, and medical imaging technologies. This research employs multivariate polynomial regression, random search, and Python-based data analysis to predict potential chemical formulas for superconductors. By systematically analyzing thousands of characterized materials and interpolating critical temperature maxima across multidimensional graphs, we identified 29 viable candidates for high-temperature superconductors and 4 families of materials that could yield high temperature superconductivity in general. Our computational approach based on extensive data in 5 parameters provides a framework for screening potential high-temperature superconductor compounds with significant technological implications. This research paper declares high temperature superconductor candidates by using supervised machine learning to optimize parameters of the superconductor. It matches the parameters at maxima of the optimization polynomial and searches for the highest T _c in the neighborhood of tested superconductors. We used random search to execute this under the restriction of a long-sustainable pressure value (1-10 atm).