A novel feature selection approach is proposed to address the curse of dimensionality and reduce the redundancy of hyperspectral data. The proposed approach is based on a new binary optimization method inspired by fractional-order Darwinian particle ...