%0 Journal Article
%T 基于贝叶斯网络的空管信息抽取相关性偏差分析<br>Bayesian Network-Based Correlation Bias Analysis for Air Traffic Control Information Extraction
%A 华炜
%A 郭庆
%A 左燕
%J Open Journal of Transportation Technologies
%P 367-378
%@ 2326-344X
%D 2024
%I Hans Publishing
%R 10.12677/ojtt.2024.135040
%X 本文首先结合事件类型识别出危险源文本中相应的事件&#65292;接着识别对相关性存在影响的中性词语&#65292;建立基于后门调整的主题分析模型&#65292;剔除主题混杂和语义相似性偏差带来的相关性分析不准确的问题。最后&#65292;基于贝叶斯网络计算不同节点的发生概率并得到非完整链路和完整链路的演化分析结果。<br>In this paper, we first identify the corresponding events in the text of the hazard source by combining the event types, then identify the neutral words that have an influence on the correlation, and establish the theme analysis model based on backdoor adjustment to eliminate the inaccuracy of the correlation analysis caused by the theme mixing and semantic similarity deviation. Finally, based on Bayesian network, we calculate the occurrence probability of different nodes and obtain the evolution analysis results of non-complete and complete links.
%K 空中交通管理&#65292
%K 安全管理&#65292
%K 贝叶斯网络模型&#65292
%K 信息抽取&#65292
%K 语义相似性<br>Air Traffic Management
%K Safety Management
%K Bayesian Network Model
%K Information Extraction
%K Semantic Similarity
%U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=97170