Publication Date



Technical Report: UTEP-CS-21-14a

Published in Julia Rayz, Victor Raskin, Scott Dick, and Vladik Kreinovich (eds.), Explainable AI and Other Applications of Fuzzy Techniques,Proceedings of the Annual Conference of the North American Fuzzy Information Processing Society NAFIPS'2021, West Lafayette, Indiana, June 7-9, 2021, Springer, Cham, Switzerland, 2022, pp. 499-504.


Many phenomena like burnout are gauged by computing a linear combination of user-provided Likert-scale values. The problem with this traditional approach is that, while it makes sense to have linear combination of weights or other physical characteristics, a linear combination of Likert-scale values like "good" and "satisfactory" does not make sense. The only reason why linear combinations are used in practice is that the corresponding data processing tools are readily available. A more adequate approach would be to use fuzzy logic -- a technique specifically designed to deal with Likert-scale values. We show that fuzzy logic actually leads to a linear combination -- but not of the original degrees, but of their transformed values. The corresponding transformation function -- as well as the coefficients of the corresponding linear combination -- must be determined from the condition that the resulting expression best fits the available data.

tr21-14.pdf (135 kB)
Original file