A rule-based named-entity recognition method for knowledge extraction of evidence-based dietary recommendations

Area of Science:

• Biomedical Informatics
• Nutritional Science
• Natural Language Processing

Background:

• Dietitians require timely access to evidence-based dietary information from scientific literature.
• Existing named-entity recognition (NER) methods primarily focus on biomedical entities like genes, proteins, and drugs.
• There is a gap in automated extraction of dietary concepts from scientific reports.

Purpose of the Study:

• To introduce drNER, a novel rule-based NER method for extracting evidence-based dietary information.
• To address the challenge of accessing and utilizing daily-arriving dietary knowledge from scientific publications.
• To be the first method specifically designed for extracting dietary concepts.

Main Methods:

• Developed drNER, a two-phase rule-based NER system.
• Phase 1: Detection and determination of dietary entity mentions.
• Phase 2: Selection and extraction of identified dietary entities.
• Evaluated drNER on text corpora from scientific websites and publications.

Main Results:

• drNER demonstrated good performance in knowledge extraction.
• The method successfully extracts evidence-based dietary recommendations.
• The system is capable of processing heterogeneous text sources.

Conclusions:

• drNER is an effective tool for extracting dietary concepts from scientific literature.
• This method facilitates dietitians' access to the latest evidence-based dietary knowledge.
• drNER represents a significant advancement in the field of dietary information extraction.