Prototypes for LOINC code predictions and imputations. Key ideas: string/text similarity metric learning, document embeddings, contrastive loss and Siamese network, classification algorithms (e.g. logistic regression, random forest), EDA via dimensionality reduction (e.g. t-SNE, PCA).

REST service that converts LOINC codes and UCUM units to a standardized representation

The Loinc Table Core