Unsupervised patient representations from clinical notes with interpretable classification decisions

2017/11/14 by Madhumita Sushil, Simon Šuster, Sushil, Madhumita +5
Biochemistry, Genetics and Molecular Biology · Computer Science · Medicine · #Biomedical Text Mining and Ontologies #Colorectal Cancer Screening and Detection #Computation and Language (cs.CL) #FOS: Computer and information sciences #Machine Learning in Healthcare

paper · pdf · doi:10.48550/arxiv.1711.05198

openalex publication_date 2017/11/14 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

We have two main contributions in this work: 1. We explore the usage of a stacked denoising autoencoder, and a paragraph vector model to learn task-independent dense patient representations directly from clinical notes. We evaluate these representations by using them as features in multiple supervised setups, and compare their performance with those of sparse representations. 2. To understand and interpret the representations, we explore the best encoded features within the patient representations obtained from the autoencoder model. Further, we calculate the significance of the input features of the trained classifiers when we use these pretrained representations as input.

Citations

• A Coefficient of Agreement for Nominal Scales
• Learning for Biomedical Information Extraction: Methodological Review of Recent Advances
• Combining LSTM and Latent Topic Modeling for Mortality Prediction

Related

• Adaptation of Autoencoder for Sparsity Reduction From Clinical Notes Representation Learning
• Effective Representations of Clinical Notes
• Improved Hierarchical Patient Classification with Language Model Pretraining over Clinical Notes
• Automated Scoring of Clinical Patient Notes using Advanced NLP and Pseudo Labeling
• Representation Learning for Electronic Health Records