Research map / Probabilistic and scientific modeling / Molecules and proteins
Precision-Recall Balanced Topic Modelling
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 |
|---|---|---|---|
| increase lambda from 0 (LDA) to 0.1 to weight precision as well as recall | topic model · NYTIMES | mean precision: -78.8 -> -72.7 | Helped |
| increase lambda from 0 (LDA) to 0.1 | topic model · NYTIMES | mean recall: -2.97 -> -3.61 | Hurt |
| increase lambda from 0 (LDA) to 0.1 | topic model · NYTIMES | topic coherence@5: -23.1 -> -16 | Helped |
| increase lambda from 0 (LDA) to 0.1 | topic model · NYTIMES | L1 distance: 1.74 -> 1.63 | Helped |
| increase lambda from 0 (LDA) to 0.1 | topic model · 20NEWSGROUP | ARI: 0.15 -> 0.209 | Helped |
| increase lambda from 0 (LDA) to 0.1 | topic model · NYTIMES | topic entropy: 5.85 -> 4.78 | Helped |
| increase lambda from 0 (LDA) to 0.1 | topic model · 20NEWSGROUP | R@10 (%): 9.85 -> 8.92 | Hurt |
| use fixed empirical distribution (lambda mixture) instead of SW document-specific learned distributions | topic model · NYTIMES | mean precision: -79.3 -> -72.7 | Helped |
| tune LDA alpha and gamma priors to trade off recall and precision (ablation) | LDA · all datasets | precision and coherence | No clear effect |
| use asymmetric topic prior instead of symmetric (ablation) | LDA · all datasets | precision | No clear effect |
| add shared background distribution to SW model (SWB) (ablation) | SW model · NYTIMES | topic quality | No clear effect |
| increase lambda from 0 on crime data (K=5) (ablation) | spatio-temporal topic model · London crime | PEI and PAI | Helped |
Closest papers by meaning
- Diagnosing and Improving Topic Models by Analyzing Posterior Variability (AAAI 2018 · similarity 0.86)
- On-the-fly Rectification for Robust Large-Vocabulary Topic Inference (ICML 2021 · similarity 0.85)
- Provable Algorithms for Inference in Topic Models (ICML 2016 · similarity 0.85)
- IMS-DTM: Incremental Multi-Scale Dynamic Topic Models (AAAI 2018 · similarity 0.85)
- On the Affinity, Rationality, and Diversity of Hierarchical Topic Modeling (AAAI 2024 · similarity 0.83)
- Provable unlearning in topic modeling and downstream tasks (ICLR 2025 · similarity 0.83)
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.