Research map / Learning theory and algorithms / Sample complexity
An Estimation and Analysis Framework for the Rasch Model
Andrew Lan, Mung Chiang, Christoph Studer
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| Use the L-MMSE estimator instead of the strongest listed nonlinear baseline | Rasch model · MT, SS, edX, and ML | ACC: 0.795 \pm 0.016 vs 0.796 \pm 0.015; 0.860 \pm 0.007 vs 0.859 \pm 0.007; 0.932 \pm 0.001 vs 0.935 \pm 0.002; 0.715 \pm 0. | No clear effect |
| Use the L-MMSE estimator instead of the strongest listed nonlinear baseline | Rasch model · MT | AUC: 0.840 \pm 0.016 vs 0.843 \pm 0.015 | Hurt |
| Use the L-MMSE estimator instead of the strongest listed nonlinear baseline | Rasch model · SS | AUC: 0.800 \pm 0.014 vs 0.803 \pm 0.013 | Hurt |
| Use the L-MMSE estimator instead of the strongest listed nonlinear baseline | Rasch model · edX | AUC: 0.900 \pm 0.004 vs 0.909 \pm 0.004 | Hurt |
| Use the L-MMSE estimator instead of the strongest listed nonlinear baseline | Rasch model · ML | AUC: 0.755 \pm 0.005 vs 0.756 \pm 0.004 | No clear effect |
| Use the L-MMSE-based upper bound rather than the Fisher information-based lower bound | PM estimator MSE analysis · Synthetic data, low SNR and small problem sizes | MSE bound tightness | Improved |
| Use the L-MMSE estimator rather than PM estimation | Rasch model · MT and ML | runtime: 0.23s and 79s; 1.9s and 528s | Improved |
Closest papers by meaning
- Fine-Grained Dynamic Framework for Bias-Variance Joint Optimization on Data Missing Not at Random (NeurIPS 2024 · similarity 0.81)
- Debiased Collaborative Filtering with Kernel-Based Causal Balancing (ICLR 2024 · similarity 0.81)
- Learning-Efficient Yet Generalizable Collaborative Filtering for Item Recommendation (ICML 2024 · similarity 0.81)
- From Pairwise to Ranking: Climbing the Ladder to Ideal Collaborative Filtering with Pseudo-Ranking (AAAI 2025 · similarity 0.80)
- Can We Trust Item Response Theory for AI Evaluation? (NeurIPS 2026 · similarity 0.80)
- Addressing Hidden Confounding with Heterogeneous Observational Datasets for Recommendation (NeurIPS 2024 · similarity 0.80)
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.
