PANC: Prototype Augmented Neighbor Constraint instance completion in knowledge graphs
作者：
Highlights：
• Instance completion task considers the rationality of prediction.
• Rely on prototype for a wider set of reasonable candidates.
• Neighbor aggregation assists to accurately discriminating knowledge triplets.
• Performance of model declines the number of neighbors exceed a threshold.
摘要
•Instance completion task considers the rationality of prediction.•Rely on prototype for a wider set of reasonable candidates.•Neighbor aggregation assists to accurately discriminating knowledge triplets.•Performance of model declines the number of neighbors exceed a threshold.
论文关键词：Knowledge graph,Completion instance,Completion neighbor,Aggregation prototype,Segmentation fuzzy,Clustering
论文评审过程：Author links open overlay panelRuixinMaabEnvelopeYunlongMaabEnvelopeHongyanZhangabBiaoMeiabGuangyueLvabEnvelopeLiangZhaoabPerson1Envelope
论文官网地址：https://doi.org/10.1016/j.eswa.2022.119013
