Accident causal chain prediction method and device based on text mining and graph neural network

• Main Authors: LI MINGHAO, QU HAOHAO, YAN WEILONG, XU XINWAN, JIN PEIJIAN, SHI ZHIKUN, GUO SHIHAO, LU XIN, LI HANG, LI CHAOQUN, SUN SHIMEI, XU JINRONG, YANG SHUO, ZHU KAI
• Format: Patent
• Language: Chinese; English
• Published: 19.03.2024
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS

The invention relates to the technical field of accident causal chain prediction, in particular to an accident causal chain prediction method and device based on text mining and a graph neural network. The invention discloses an accident causal chain prediction method based on text mining and a graph neural network. The method comprises the following steps: obtaining an entity lexicon and a reason label lexicon according to an original text; performing graph construction according to the entity lexicon and the reason label lexicon to obtain an accident causal chain; performing model construction according to the accident causal chain to obtain a graph attention network model; based on the original text and a preset expert accident causal chain, optimizing the graph attention network model to obtain an optimized graph attention network model; obtaining a target accident text; and according to the target accident text, performing prediction through the optimized graph attention network model to obtain a target