Stabilizing High-Dimensional Prediction Models Using Feature Graphs

We investigate feature stability in the context of clinical prognosis derived from high-dimensional electronic medical records. To reduce variance in the selected features that are predictive, we introduce Laplacian-based regularization into a regression model. The Laplacian is derived on a feature graph that captures both the temporal and hierarchic relations between hospital events, diseases, and interventions. Using a cohort of patients with heart failure, we demonstrate better feature stability and goodness-of-fit through feature graph stabilization.

Authors

• Shivapratap Gopakumar
• Truyen Tran Applied Artificial Intelligence Institute, Deakin University
• Tu Dinh Nguyen
• Dinh Phung
• Svetha Venkatesh Deakin University

Keywords

• Stability criteria
• Indexes
• Predictive models
• Data models
• Heart
• Feature extraction
• Prediction Model
• Graph Features
• Heart Failure
• Electronic Health Records
• Temporal Relationship
• Intersection Over Union
• Stable Characteristics
• Consistency Index
• Feature Subset
• Events In Group
• Correlated Features
• Hosmer-Lemeshow Test
• Feature Weights
• Elastic Net
• Top Features
• Elastic Net Regression
• Resampled Data
• Electronic Medical Record Data
• Clinical Prediction Models
• Regular Graphs
• Laplacian Regularization
• Lasso Regularization
• Time Granularity
• Automatic Feature Selection
• Electronic Medical Record Database
• Chronic Ischemia
• Multiple Time Periods
• Validation Cohort
• Ridge Regression
• Feature Selection Algorithm
• Biomedical computing
• electronic medical records
• stability
• Aged
• Female
• Humans
• Male
• Models, Biological
• Models, Statistical
• Reproducibility of Results
• Risk Factors