講演情報
[3D-01]Semantic-Driven Anomaly Analysis and Root Cause Identification in Heterogeneous Sensor Data
*Berjab Nesrine1 (1. 東京科学大学)
発表者区分：一般
論文種別：ロングペーパー
インタラクティブ発表：なし
論文種別：ロングペーパー
インタラクティブ発表：なし
キーワード：
IoT、Sensor networks、Semantic technologies、Anomaly detection、Root cause analysis
This paper introduces the Anomaly Analysis on Semantic Sensor (A2S2) ontology, an RDF-based framework for modeling and analyzing sensor data alongside heterogeneous contextual information. The proposed ontology extends the Sensor, Observation, Sample, and Actuator (SOSA) ontology to enable comprehensive analysis of sensor data within specific temporal and spatial contexts, facilitating the identification of root causes for observed anomalies. By integrating semantic techniques such as SPARQL queries and symbolic reasoning, the framework supports the automatic derivation of complex physical process models. Experimental evaluations demonstrate the efficacy of the approach, achieving high accuracy in identifying root causes of anomalies. This work provides a foundation for robust anomaly detection in heterogeneous sensor networks.