Grounded theory methodology for knowledge engineering

Grounded theory methods of inductive qualitative field research have been tried and tested in the social sciences with their strong commitment to empirical techniques. They can be used in knowledge engineering for grounding systems on empirical evidence wherever the domain experts’ natural way of thinking is qualitative. After an introductory explanation of grounded theory development we go through two recent studies that discover strategies – those of companies redressing their economic situation and those of expert summarizers doing their job. The work about expert summarizing is pertinent to empirical AI. Therefore I discuss it in detail. The resulting model has given rise to a small implemention. It demonstrates how expert summarization works and that grounding a summarization system on human cognitive processing should be feasible. Grounded theories may be of general interest to AI colleagues who want their systems to have an empirical foundation.