Machine learning yields new candidates for layered superconductors

Maximum difference of electronegativity appears to play major role when using molecular attributes to find potential novel superconconducting materials over a wide range of critical temperatures.