Real-time wildfire detection with semantic explanations
作者：
Highlights：
• Our first explainable CEP framework can process thousands of events per second.
• Predictive predicates can be learned within seconds.
• Our framework overcomes the accuracy-explainability tradeoff with 0.93 F1-score.
• Our framework (FADE) can discover optimal explanations with nearly 0.9 accuracy.
• Our explanation predicates can cover more than 80% of true predictive features.
摘要
•Our first explainable CEP framework can process thousands of events per second.•Predictive predicates can be learned within seconds.•Our framework overcomes the accuracy-explainability tradeoff with 0.93 F1-score.•Our framework (FADE) can discover optimal explanations with nearly 0.9 accuracy.•Our explanation predicates can cover more than 80% of true predictive features.
论文关键词：Wildfire detection,Real-time monitoring,Explanation discovery,Complex event processing
论文评审过程：Received 29 October 2021, Revised 24 March 2022, Accepted 26 March 2022, Available online 9 April 2022, Version of Record 21 April 2022.
论文官网地址：https://doi.org/10.1016/j.eswa.2022.117007
