Explanation-Guided Medical Named Entity Recognition with Stability and Boundary Awareness for Atopic Dermatitis

The study introduces a stability and boundary-aware explanation-guided NER framework specifically for Chinese atopic dermatitis clinical texts, enhancing the robustness of medical named entity recognition. Utilizing perturbation-based analysis and an adaptive fusion strategy, the framework improves explanation stability and entity boundary sensitivity, leading to consistent performance gains across various NER models. This advancement is significant for practitioners as it integrates reliable explanation signals into NER training, supporting improved clinical decision-making and medical knowledge applications.