Optimizing Botanical Data Integrity: A Comparative Study of Text Similarity Methods

In this study, we address the challenges of managing authorship nomenclature as dictated by the International Code of Nomenclature for algae, fungi, and plants (ICN), within the Begoniaceae and Bignoniaceae families databases. Our goal was to evaluate various text similarity algorithms for their effectiveness in deduplicating botanical data, ensuring accuracy in authorship and synonymy. Our results highlighted Smith-Waterman’s superior balance in precision, recall, and F1 Score, suggesting its potential as a robust solution for improving database integrity. The study also demonstrates the importance of fine-tuning these algorithms to navigate the unique challenges of botanical data management, emphasizing the necessity for specialized approaches in this field.