Case-Based Reasoning via Comparing the Strength Order of Features
English
Yu, Liuwen[University of Luxembourg > Interdisciplinary Centre for Security, Reliability and Trust (SNT) > PI VDT]
Gabbay, Dov M.[University of Luxembourg > Faculty of Science, Technology and Communication (FSTC) > Computer Science and Communications Research Unit (CSC) > ; King's College, London]
Springer Science and Business Media Deutschland GmbH
13283 LNAI
143-151
Yes
International
03029743
4th International Workshop on Explainable and Transparent AI and Multi-Agent Systems, EXTRAAMAS 2022
9 May 2022 through 10 May 2022
[en] Argumentation ; Case-based reasoning ; Legal reasoning ; Strength of features ; Computers ; Case based ; Casebased reasonings (CBR) ; New case ; Relative strength ; Similar case ; Strength of feature ; Case based reasoning
[en] Case-based reasoning (CBR) is broadly speaking a method of giving a verdict/decision on a new case query by comparing it with verdicts/decisions of known similar cases. Similarity of cases is determined either by best distance of the query case from the known cases and recently also using argumentation. The approach of this paper is not to rely on similarity or argumentation, but to use the entire set of known cases and their known verdicts to define the relative strength and importance of all the features involved in these cases. We then decide the verdict for the new case based on the strength of the features appearing in it.