Model-agnostic and diverse explanations for streaming rumour graphs
作者：
Highlights：
• Our approach yields explanations of much higher accumulated utility up to 16x.
• Our approach is robust to concept drift with a 67.95% of explanation accuracy.
• Our approach has indexing performance significantly better in time and space.
• Our embedding-based similarity is much faster to compute than graph measures.
• Our example-based explanations also support the detection accuracy (>=0.95).
摘要
•Our approach yields explanations of much higher accumulated utility up to 16x.•Our approach is robust to concept drift with a 67.95% of explanation accuracy.•Our approach has indexing performance significantly better in time and space.•Our embedding-based similarity is much faster to compute than graph measures.•Our example-based explanations also support the detection accuracy (>=0.95).
论文关键词：Explainable rumour detection,Data stream,Social networks,Graph embedding
论文评审过程：Received 25 May 2022, Revised 29 June 2022, Accepted 11 July 2022, Available online 22 July 2022, Version of Record 29 July 2022.
论文官网地址：https://doi.org/10.1016/j.knosys.2022.109438
