Open AccessElicitation of knowledge from a defence expert
Author(s) -
Sarah O’Brien,
W. G. Proud,
Margaret Wilson
Publication year - 2020
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1507/10/102027
Subject(s) - subject matter expert , expert elicitation , expert system , relation (database) , computer science , field (mathematics) , key (lock) , task (project management) , legal expert system , sort , domain (mathematical analysis) , data science , knowledge management , artificial intelligence , information retrieval , data mining , engineering , mathematics , statistics , mathematical analysis , computer security , systems engineering , pure mathematics
The aim of this work is to understand the way that a defence expert defines the concept of importance in relation to the ideas contained in a scientific document. The expert’s views on the importance of the concepts in this document were elicited in two phases. In the first phase, the expert was asked to summarise an eight-page document on the effects of electromagnetic fields on propellant combustion. Completion of this task generated a series of ’key points’. Phase two of the methodology was a sit-down interview with the expert. This interview comprised three parts: asking the expert to talk through why each of the key points were important, asking the expert to sort the key points into categories according to how important they are and then asking the expert to generate categories of why the points are important. The techniques used for expert elicitation proved highly successful in relation to this domain of knowledge. Not only were the procedures able to extract the underlying categories through which the expert structured their understanding of the field, but the results indicated reliability in the content of knowledge extracted through different methods. Subsequent papers in this project compare this work to parallel analysis conducted using Natural Language Processing tools.