基于图嵌入的智能建筑异常数据检测

Anomaly Detection in Intelligent Building Management System Using Graph Embedding

智能建筑系统中各种传感器生成了大量数据,数据管理和分析面临极大挑战,而设备异常检测是数据管理和分析的关键任务。为解决这个问题,本文提出了一种基于图嵌入的异常检测方法,通过将建筑设备数据转换为图表示,并将其嵌入到低维空间中,可以有效地检测建筑设备中的异常。该方法结合了图嵌入和深度学习,具有良好的线性可扩展性,并能准确地捕捉建筑物中不同传感器之间的复杂非线性关系。本文还提供了研究案例,进一步说明所提出的基于图嵌入的异常检测方法在实际应用中的有效性。

Intelligent building systems generate a large amount of data from various sensors,which poses great challenges for management and analysis.However,anomaly detection of those devices is a key task for managing and analyzing these data.To solve this problem,this paper proposes an anomaly detection method based on knowledge graph embedding.By converting the collected data into a graph representation and embedding it into a low-dimensional space,it can effectively detect anomalies in building equipment.This method combines graph embedding and deep learning,has good linear scalability,and can accurately capture complex nonlinear relationships between different sensors in building management system.This paper also provides a case study that further demonstrates the effectiveness of the proposed graph embedding-based anomaly detection method in practical applications.