During knowledge elicitation or expert labelling human experts may be asked to assess their level of confidence in a rule, argument or classification. This confidence rating may be a numerical scale, such as 0-10, a percentage, or more qualitatively in a Likert scale.
Used in Chap. 18: page 303