Benchmarking of Sentence Transformers in Text-Based Drug-Drug Interaction Classification
26th International Conference on Bioinformatics and Computational Biology, BIOCOMP 2025, 11th International Conference on Biomedical Engineering and Sciences, BIOENG 2025, and 11th International Conference on Health Informatics and Medical Systems, HIMS 2025, Held as Part of the World Congress in Computer Science, Computer Engineering, and Applied Computing, CSCE 2025, Nevada, Amerika Birleşik Devletleri, 21 - 24 Temmuz 2025, cilt.2935 CCIS, ss.235-248, (Tam Metin Bildiri)
- Yayın Türü: Bildiri / Tam Metin Bildiri
- Cilt numarası: 2935 CCIS
- Doi Numarası: 10.1007/978-3-032-22199-5_17
- Basıldığı Şehir: Nevada
- Basıldığı Ülke: Amerika Birleşik Devletleri
- Sayfa Sayıları: ss.235-248
- Anahtar Kelimeler: Classification of Drug-Drug Interaction, Neural Network, Sentence Transformer Model
- Ankara Üniversitesi Adresli: Evet
Özet
The detection of drug-drug interactions (DDIs) has become a key research focus due to its significant impact on healthcare and drug safety. Various machine learning methods have been proposed for DDI identification and classification, while most studies focus on drug chemical properties, with few considering the impact of textual data. This study investigates these influencing factors to assess their contributing and compromising effects on DDI identification and classification. To this end, the impact of chemical features and the most accurate representation of the textual data has also been assessed, and several experiments have been performed to obtain a fair comparison. The results indicated that text embeddings significantly enhanced DDI classification capabilities compared with models relying solely on chemical features.