MedSkim: Denoised Health Risk Prediction via Skimming Medical Claims Data

Keyphrases

• EHR Data 100%
• Medical Claims Data 100%
• Health Risk Prediction 100%
• Skipping 50%
• Remove Noise 33%
• Code-based 16%
• Hidden States 16%
• Risk Prediction 16%
• Source Code 16%
• Disease Conditions 16%
• Cohen's kappa 16%
• Coarse-grained Representation 16%
• Diagnostic Codes 16%
• Individual Diagnosis 16%
• Representation Learning 16%
• Two-module 16%
• Regularization Term 16%
• Ablation Study 16%
• Prediction Module 16%

Computer Science

• Recurrent Neural Network 100%
• Case Study 33%
• Experimental Result 33%
• Representation Learning 33%
• Regularization Term 33%
• Disease Condition 33%