🏥 Clinical Disease Entity Extractor

Model: ugaray96/biobert_ncbi_disease_ner  |  Base: BioBERT  |  Trained on: NCBI Disease dataset

Paste any clinical free-text below. The model will identify disease entities and return a confidence score for each.

Confidence Score: Average softmax probability across entity tokens. Scores below 50% indicate the model is not plurality-confident in its DISEASE prediction — a natural decision boundary for a 3-class problem — a useful signal for monitoring model reliability in production.

Custom model trained on MedMentions using emilyalsentzer/Bio_ClinicalBERT available at acebirim/snomed-ner-model. Built as part of an NHS/OneLondon Secure Data Environment capstone project.