Answers are grounded in this paper and cite their sources.
If you wanted to make some money out of natural language processing, an appropriate strategy might be to identify an area of technology that was relatively mature—one where the more fundamental technical problems had been resolved through a significant amount of research activity—and then identify potential applications for that technology.
No takes yet. Share an insight, caveat, or question.
Robert Dale (2004) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: