Personalized medicine requires the ability to create precision medicine based on the personal health records (EHR, EMR).
Problem: Customer needed a software platform for clinical concept extraction from patient case notes.
Methodology: Cenacle built an innovative solution based on the below methodology:
- Natural Language Processing (NLP) to identify: diseases, symptoms, medications, procedures etc.
- Deep Learning Artificial Intelligence (AI) architecture using RNN with LSTM for sequence labelling
- Train the model on annotated text
- Classifier built based on UMLS and MIMIC data set models
- UMLS: F1 Score: 0.788
- MIMIC: F1 Score: 0.725