Code and Data for 'Hierarchical Entity Disambiguation for Aviation Maintenance Knowledge Graph Construction: A Multi-Relational CRF Approach'
Year: 2026
Type: software
Abstract: This deposit accompanies the IEEE Access manuscript "Hierarchical Entity Disambiguation for Aviation Maintenance Knowledge Graph Construction: A Multi-Relational CRF Approach" (Manuscript ID Access-2026-34396). It provides the complete code and data required to reproduce the entity-linking and named-entity-recognition (NER) experiments. Contents: - 04_Implementation_Code/ — Python implementation (BiLSTM-CRF NER model, canopy-based candidate indexing, multi-relational CRF entity disambiguation, and the string-matching, TF-IDF, and independent-CRF baselines) - 04_Implementation_Code/models/ner_model.pt — trained BiLSTM-CRF NER checkpoint - 03_Data_Outputs/eval_data/ — evaluation data (73-document test split, 209 gold links, knowledge base, BIO-labeled NER test split) - 03_Data_Outputs/ — reproduced result files and the integrity manifest Dataset split: 336 training, 72 development, 73 test documents. Reproduction commands are documented in README.md.
Author: Anonymous, Author
Language: English
Topic: Topic Modeling
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7214087859
