Safe medicine recommendation via star interactive enhanced-based transformer model

Nanxin Wang, Xiaoyan Cai, Libin Yang, Xin Mei

Abstract: Highlights•A novel neural model, called SIET, is proposed for medicine recommendation.•An inductive learning-based method is developed to learn entity (disease, medicine) embedding.•SIET achieves state-of-art performance in MIMIC-III, DrugBank and ICD-9 ontology datasets.