Random Forest Regressor based superconductivity materials investigation for critical temperature prediction

Ever since its invention over past hundred years, superconductivity has been the subject of intense investigation. However, numerous aspects of this unusual phenomenon stay unknown, the most notable of which being the relationship among superconductivity, compound/structural assets of materials as well. Every superconductor materials transition temperature that lies in between 1 Kelvin and 10 Kelvin. Based on critical temperature of materials, superconductivity materials classified into two namely less than 10 Kelvin, greater than 10 Kelvin. Several regression models are developed here to analyze the critical temperatures of more than 12,000 known superconductors accessible through Super Con metadata, in order to sustain. After studying and implementing the aforementioned techniques, Random Forest Regressor stood out and gave the best results in terms of R^2 score metrics initial value as 91.2% and after normalizing features in superconductivity metadata, R^2 score value reaches 92.79% in predicting the temperature values of superconductors.