Naslov
A Comparative Study of Defeasible Argumentation and Non-monotonic Fuzzy Reasoning for Elderly Survival Prediction Using Biomarkers

Autori
Rizzo, Lucas ; Majnaric, Ljiljana ; Longo, Luca

Izvornik
AI*IA 2018 – Advances in Artificial Intelligence (0302-9743) LNCS 11298 (2018); 197-209

Vrsta, podvrsta i kategorija rada
Radovi u časopisima, članak, znanstveni

Ključne riječi
Argumentation Theory, Non-monotonic Reasoning, Defea-sible Reasoning, Fuzzy Reasoning, Possibility Theory, Biomarkers

Sažetak
Computational argumentation has been gaining momentum as a solid theoretical research discipline for inference under uncertaint ywith incomplete and contradicting knowledge. However, its practical counter part is under developed, with a lack of studies focused on the investigation of its impact in real-world settings and with real knowl-edge. In this study, computational argumentation is compared against non- monotonic fuzzy reasoning and evaluated in the domain of biological markers for the prediction of mortality in an elderly population. Differ- ent non-monotonic argument-based models and fuzzy reasoning modelshave been designed using an extensive knowledge base gathered from an expert in the field. An analysis of the true positive and false positive rate of the inferences of such models has been performed. Findings indicate a superior inferential capacity of the designed argument-based models.

Izvorni jezik
Engleski

POVEZANOST RADA