.png){kind=icon}
.png){kind=icon}
Return
Ontology-Based Forest Fire Management Using Complex Event Processing and Large Language Models
.png){kind=icon}
PRE
AI
DOI：10.1007/978-3-032-02049-9_7.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Forests sustain ecology, but wildfires are a significant threat. Fire weather indices help assess hazards but require constant realtime processing. To address this, we developed a Semantic Sensor Network (SSN) ontology model using data from Monesterial Natural Park, enhanced with Semantic Web Rules Language (SWRL). The system integrates Large Language Models (LLMs) and Complex Event Processing (CEP) engines for real-time fire detection. Sensor networks collect climate data (humidity, temperature, wind speed, etc.), which is processed via Spark Streaming and CEP to identify fire-related events. LLMs analyze detected events, while SPARQL queries retrieve relevant insights from the ontology. The results are combined to estimate the overall risk, allowing informed decision-making within a comprehensive Decision Support System (DSS) framework. It makes it easier to understand and deal with the risks of wildfires, as shown by tests that use ontology metrics, query-based testing, event alerts, and LLM performance (F1 score, precision, and recall).
Keywords:
LLMs
SSN
Spark
CEP
Journal
D
IF:
0
Papers:
25
Citations:
0