It is mentioned here that
It is possible to correct for the DFT band gap error by introducing an empirical scissors correction, which is effectively a rigid shift of the conduction band with respect to the valence band. This approach produces satisfactory results for such properties as optical spectra, provided the experimental band gap is accurate. The scissors operator scheme is very difficult to use in predictive studies and in the context of band gap engineering, when there is little or no experimental information about the electronic structure
How can I implement scissor correction in VASP to study optical properties?