This repository contains research and implementation related to the paper:

Semantic Modeling of ECU 911 Emergencies Using NLP and Ontologies

Published in Revista Tecnológica ESPOL (RTE).

DOI: https://doi.org/10.37815/rte.v37nE1.1351

Emergency call centers such as ECU 911 receive large volumes of unstructured textual data from citizens.
This project proposes a hybrid approach that combines Natural Language Processing (NLP) and semantic technologies to transform raw emergency reports into structured knowledge.

The system integrates:

• Named Entity Recognition (NER)
• Semantic classification of emergency incidents
• Ontology-based knowledge representation
• Logical inference using SWRL rules

The proposed pipeline consists of the following components:

1. NLP Processing

   • BERT model for Named Entity Recognition (NER)
   • XLM-RoBERTa for zero-shot emergency classification

2. Data Processing

   • Cleaning and normalization of emergency reports
   • Extraction of semantic features

3. Knowledge Representation

   • OWL ontology developed in Protégé
   • RDF knowledge graph generation

4. Inference Layer

   • SWRL rules derived from decision trees
   • Semantic reasoning using Pellet reasoner

5. Validation

   • SPARQL and SQWRL queries
   • Comparison with real ECU 911 incident labels

The system achieved 96.7% accuracy when inferring the priority level of emergency incidents.

Python
Transformers (HuggingFace)
BERT
XLM-RoBERTa
Protégé
OWL
SWRL
SPARQL
Prolog

Paltin, D., Mejía, J., Orellana, M., & Zambrano-Martinez, J. (2025).
Semantic Modeling of ECU 911 Emergencies Using NLP and Ontologies.
Revista Tecnológica ESPOL.

https://doi.org/10.37815/rte.v37nE1.1351

Danny Leonardo Paltin