# Phenotype Inference with Semi-Supervised Mixed Membership Models

## TL;DR

This paper introduces SS3M, a semi-supervised probabilistic model that efficiently learns disease phenotypes from clinical data with minimal labeled examples, improving interpretability and disease specificity.

## Contribution

The paper presents SS3M, a novel semi-supervised mixed membership model that addresses the limitations of existing phenotyping methods by requiring fewer labels and producing interpretable disease phenotypes.

## Key findings

- SS3M effectively learns disease-specific phenotypes from limited labeled data.
- The model produces interpretable phenotypes aligned with clinical characteristics.
- SS3M outperforms purely supervised or unsupervised methods in phenotype quality.

## Abstract

Disease phenotyping algorithms process observational clinical data to identify patients with specific diseases. Supervised phenotyping methods require significant quantities of expert-labeled data, while unsupervised methods may learn non-disease phenotypes. To address these limitations, we propose the Semi-Supervised Mixed Membership Model (SS3M) -- a probabilistic graphical model for learning disease phenotypes from clinical data with relatively few labels. We show SS3M can learn interpretable, disease-specific phenotypes which capture the clinical characteristics of the diseases specified by the labels provided.