Research map / Probabilistic and scientific modeling / Variational and generative sampling
Ordinal Non-negative Matrix Factorization for Recommendation
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 |
|---|---|---|---|
| Use ordinal ratings instead of binarizing them | Recommendation on MovieLens · MovieLens, NDCG@100 thresholds s=1, 4, 6, 8 | NDCG@100: 0.433 -> 0.444 (s=1); 0.430 -> 0.444 (s=4); 0.421 -> 0.439 (s=6); 0.408 -> 0.414 (s=8) | Helped |
| Use ordinal ratings instead of binarizing them | Recommendation on MovieLens · MovieLens, NDCG@100 threshold s=10 | NDCG@100: 0.369 -> 0.353 | Hurt |
| Use quantized ordinal play counts instead of binarized play counts | Recommendation on Taste Profile · Taste Profile, NDCG@100 thresholds s=1 and s=3 | NDCG@100: 0.210 -> 0.213 (s=1); 0.173 -> 0.174 (s=3) | Helped |
| Use quantized ordinal play counts instead of raw counts | Recommendation on Taste Profile · Taste Profile, NDCG@100 thresholds s=6, 11, 21, 51 | NDCG@100: 0.154 -> 0.153 (s=6); 0.137 -> 0.135 (s=11); 0.128 -> 0.123 (s=21); 0.121 -> 0.117 (s=51) | Hurt |
| Use quantized ordinal play counts instead of raw counts | Recommendation on Taste Profile · Taste Profile, test non-zero entries | log-likelihood: -3.0 \cdot 10^{5} -> -2.8 \cdot 10^{5} | Helped |
| Set q(n_ui) to a point mass at 1 for positive observations (ablation) | IG-OrdNMF inference · Recommendation | recommendation performance | Hurt |
Closest papers by meaning
- Coupled Poisson Factorization Integrated With User/Item Metadata for Modeling Popular and Sparse Ratings in Scalable Recommendation (AAAI 2018 · similarity 0.88)
- Indexable Probabilistic Matrix Factorization for Maximum Inner Product Search (AAAI 2016 · similarity 0.88)
- Hierarchical Negative Binomial Factorization for Recommender Systems on Implicit Feedback (AAAI 2021 · similarity 0.87)
- Hierarchical Compound Poisson Factorization (ICML 2016 · similarity 0.87)
- Recommendations with Sparse Comparison Data: Provably Fast Convergence for Nonconvex Matrix Factorization (ICML 2025 · similarity 0.87)
- Preference Diffusion for Recommendation (ICLR 2025 · similarity 0.85)
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.