New xMICD method enhances interpretability of patient diagnosis representations

Researchers have developed xMICD, a novel method for creating interpretable low-dimensional representations of patient diagnoses from International Classification of Diseases (ICD) codes. This approach balances predictive performance with clinical interpretability by integrating diagnostic groupings with similarity in a pre-trained ICD embedding space. Experiments show xMICD achieves comparable predictive accuracy to existing embedding-based methods like ICD2Vec on clinical prediction tasks, while maintaining feature interpretability tied to recognizable diagnostic groups. AI

IMPACT This method could improve the interpretability of machine learning models used in healthcare, facilitating better clinical decision-making.

RANK_REASON The cluster contains a research paper detailing a new methodology for representing medical data. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• electronic health records
• ICD2Vec
• ICD codes
• International Statistical Classification of Diseases and Related Health Problems
• xMICD

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →