An NLP framework for extracting causes, consequences, and hazards from occurrence reports to validate a HAZOP study | IEEE Conference Publication | IEEE Xplore