Research map / Time series / Causal and generative time series
Learning of Cluster-based Feature Importance for Electronic Health Record Time-series
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add clustering component (CAMELOT encoder-identifier-predictor) vs encoder-predictor only (ablation) | CAMELOT · HAVEN | AUROC: 0.73 vs 0.57 | Helped |
| add clustering component (vs encoder-predictor only) (ablation) | CAMELOT · MIMIC | AUROC: 0.72 vs 0.65 | Helped |
| replace distribution loss Ldist with AC-TPC sample-wise entropy loss (ablation) | CAMELOT · HAVEN | AUROC: 0.73 vs 0.67 | Helped |
| replace distribution loss Ldist with AC-TPC sample-wise entropy loss (ablation) | CAMELOT · MIMIC | AUROC: 0.72 vs 0.69 | Mixed |
| remove distribution loss Ldist (alpha=0) (ablation) | CAMELOT · HAVEN | AUROC: 0.73 vs 0.65 | Helped |
| remove distribution loss Ldist (alpha=0) (ablation) | CAMELOT · MIMIC | AUROC: 0.70 vs 0.72 | No clear effect |
| remove cluster separation loss Lclus (beta=0) (ablation) | CAMELOT · HAVEN | AUROC: 0.73 vs 0.65 | Helped |
| remove cluster separation loss Lclus (beta=0) (ablation) | CAMELOT · MIMIC | AUROC: 0.70 vs 0.72 | No clear effect |
| remove both Ldist and Lclus (alpha=beta=0) (ablation) | CAMELOT · HAVEN | AUROC: 0.73 vs 0.61 | Helped |
| remove both Ldist and Lclus (alpha=beta=0) (ablation) | CAMELOT · MIMIC | AUROC: 0.72 vs 0.67 | Helped |
| remove feature-time attention layer (ablation) | CAMELOT · HAVEN | AUROC: 0.69 vs 0.73 (F1 and recall equal to full model) | No clear effect |
| remove feature-time attention layer (ablation) | CAMELOT · MIMIC | AUROC: 0.72 vs 0.65 | Helped |
Closest papers by meaning
- Temporal Phenotyping using Deep Predictive Clustering of Disease Progression (ICML 2020 · similarity 0.92)
- Risk Horizons: Structured Hypothesis Spaces for Longitudinal Clinical Prediction (NeurIPS 2026 · similarity 0.85)
- Exploiting Negative Samples: A Catalyst for Cohort Discovery in Healthcare Analytics (ICML 2024 · similarity 0.85)
- NeuralCohort: Cohort-aware Neural Representation Learning for Healthcare Analytics (ICML 2025 · similarity 0.85)
- ACES: Automatic Cohort Extraction System for Event-Stream Datasets (ICLR 2025 · similarity 0.84)
- SMART: Towards Pre-trained Missing-Aware Model for Patient Health Status Prediction (NeurIPS 2024 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.