Improvements proposed to noisy-OR derivatives for multi-causal analysis: A case study of simultaneous electromagnetic disturbances

Simple item page
dc.contributor.authorDevaraj, Lokesh
dc.contributor.authorKhan, Qazi Mashaal
dc.contributor.authorRuddle, Alastair R.
dc.contributor.authorDuffy, A. P.
dc.contributor.authorPerdriau, Richard
dc.contributor.authorKoohestani, Mohsen
dc.date.acceptance2023-10-24
dc.date.accessioned2024-09-04T09:52:09Z
dc.date.available2024-09-04T09:52:09Z
dc.date.issued2023-11-09
dc.description.abstractIn multi-causal analysis, the independence of causal influence (ICI) assumed by the noisy-OR (NOR) model can be used to predict the probability of the effect when several causes are present simultaneously, and to identify (when it fails) inter-causal dependence (ICD) between them. The latter is possible only if the probability of observing the multi-causal effect is available for comparison with a corresponding NOR estimate. Using electromagnetic interference in an integrated circuit as a case study, the data corresponding to the probabilities of observing failures (effect) due to the injection of individual (single cause) and simultaneous electromagnetic disturbances having different frequencies (multiple causes) were collected. This data is initially used to evaluate the NOR model and its existing derivatives, which have been proposed to reduce the error in predictions for higher-order multi-causal interactions that make use of the available information on lower-order interactions. Then, to address the identified limitations of the NOR and its existing derivatives, a new deterministic model called Super-NOR is proposed, which is based on correction factors estimated from the available ICD information.
dc.exception.reasonnot deposited within three months of publication
dc.funderEuropean Union (EU) Horizon 2020
dc.identifier.citationDevaraj, L. et al. (2024) Improvements proposed to noisy-OR derivatives for multi-causal analysis: A case study of simultaneous electromagnetic disturbances. International Journal of Approximate Reasoning, 164, 109068
dc.identifier.doihttps://doi.org/10.1016/j.ijar.2023.109068
dc.identifier.issn0888-613X
dc.identifier.urihttps://hdl.handle.net/2086/24170
dc.peerreviewedYes
dc.publisherElsevier
dc.relation.ispartofInternational Journal of Approximate Reasoning
dc.titleImprovements proposed to noisy-OR derivatives for multi-causal analysis: A case study of simultaneous electromagnetic disturbances
dc.typeArticle
oaire.citation.volume164

Files

License bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
license.txt
Size:
4.2 KB
Format:
Item-specific license agreed upon to submission
Description:
Download

Collections

School of Engineering and Sustainable Development