A method for logical anomaly discovery based on synergistic integration of knowledge constraints and operational data with application in civil aviation ticketing
MENG Fan1, CHEN Bing2, ZHU Shaohong3, YAO Shizhou2
Abstract：The business rules governing civil aviation ticket pricing are intricate， and the logic underlying these rules evolves rapidly. Traditional unsupervised anomaly detection methods fall short in detecting business logic defects， often leading to revenue losses by the time anomalies are identified post-factum. To address this， this paper introduces a logic anomaly detection method that integrates knowledge constraints with operational data collaboration and applies it to the civil aviation ticket sales scenario. The proposed method comprises three modules. 1） Anchor Selection Strategy Module， which identifies anchor business rule sets and anchor operational data sets through a strategy based on high-frequency and persistent route characteristics. 2） Feature Discrimination Enhancement Module， which constructs a comparative dataset by combining the anchor business rule set and anchor operational data set， and employs comparative learning to enhance the feature discrimination of operational data. 3） Anomaly Matching Degree Discovery Module， which utilizes a feedforward neural network to assess the matching degree between business rules and operational data， designating log data with low matching degrees as logic anomalies. This paper constructs a real-world civil aviation ticket sales dataset and validates the proposed method from three perspectives： detection performance， ablation， and operational efficiency scalability. Experimental results demonstrate that the method can significantly improve anomaly detection performance by enhancing the feature discrimination of operational data.
孟凡,陈兵,朱少红,姚世洲. 基于知识约束和运行数据协同融合的逻辑异常发现方法及民航客票应用[J]. 华中师范大学学报(自然科学版), 2025, 59(5): 669-676.
MENG Fan,CHEN Bing,ZHU Shaohong,YAO Shizhou. A method for logical anomaly discovery based on synergistic integration of knowledge constraints and operational data with application in civil aviation ticketing. journal1, 2025, 59(5): 669-676.
2025, Vol. 59 
