TY - GEN
T1 - Selection of T-Norms for Calculating Belief Measure and Their Influence on Support Decision with Uncertainty
AU - Porębski, Sebastian
N1 - Publisher Copyright:
© 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.
PY - 2022
Y1 - 2022
N2 - The work deals with information processing with belief function of Dempster-Shafer theory (DST) generalized for a fuzzy environment. Within the domain of explainable artificial intelligence, different set theories are proposed that try to upgrade the system readability while keeping its reliability: fuzzy, intuitionistic, rough, soft to name a few. Usually they are connected to particular knowledge-based inference mechanisms: probabilistic, evidential, possibilistic, neuro-fuzzy reasoning. This chapter focuses on effective and verified decision support solution which origins from DST and allows processing uncertain knowledge and imprecise information. In this approach, where aforementioned sources of knowledge are described in a fuzzy form, accurate measure of their correspondence needs to be examined. Hence, study validates the method of fuzzy set inclusion measure calculation, particularly exploring different triangle operations. Comparison of experimental differences reveals that minimum operation can be outperformed by other t-norms and their choose is highly problem-depended. The final results let us to increase quality of decision support systems in their further development.
AB - The work deals with information processing with belief function of Dempster-Shafer theory (DST) generalized for a fuzzy environment. Within the domain of explainable artificial intelligence, different set theories are proposed that try to upgrade the system readability while keeping its reliability: fuzzy, intuitionistic, rough, soft to name a few. Usually they are connected to particular knowledge-based inference mechanisms: probabilistic, evidential, possibilistic, neuro-fuzzy reasoning. This chapter focuses on effective and verified decision support solution which origins from DST and allows processing uncertain knowledge and imprecise information. In this approach, where aforementioned sources of knowledge are described in a fuzzy form, accurate measure of their correspondence needs to be examined. Hence, study validates the method of fuzzy set inclusion measure calculation, particularly exploring different triangle operations. Comparison of experimental differences reveals that minimum operation can be outperformed by other t-norms and their choose is highly problem-depended. The final results let us to increase quality of decision support systems in their further development.
KW - Approximate reasoning
KW - Fuzzy inclusion measure
KW - Fuzzy sets
KW - Interpretable decision support systems
UR - https://www.scopus.com/pages/publications/85126225519
U2 - 10.1007/978-3-030-95929-6_18
DO - 10.1007/978-3-030-95929-6_18
M3 - Conference contribution
AN - SCOPUS:85126225519
SN - 9783030959289
T3 - Lecture Notes in Networks and Systems
SP - 229
EP - 240
BT - Uncertainty and Imprecision in Decision Making and Decision Support
A2 - Atanassov, Krassimir T.
A2 - Atanassova, Vassia
A2 - Kacprzyk, Janusz
A2 - Kałuszko, Andrzej
A2 - Krawczak, Maciej
A2 - Owsiński, Jan W.
A2 - Sotirov, Sotir S.
A2 - Sotirova, Evdokia
A2 - Szmidt, Eulalia
A2 - Zadrożny, Sławomir
PB - Springer Science and Business Media Deutschland GmbH
T2 - 16th National Conference on Operationaland Systems Research, BOS-2020 and 19th International Workshop on Intuitionistic Fuzzy Sets and Generalized Nets, IWIFSGN-2020
Y2 - 14 December 2020 through 15 December 2020
ER -