AI Paper Atlas

Research map / Probabilistic and scientific modeling / Variational and generative sampling

A Statistical Approach for Optimal Topic Model Identification

Craig M. Lewis, Francesco Grossetti

JMLR 2022 · Probabilistic and scientific modeling › Variational and generative sampling · approach: Tractable probabilistic models · 0 citations (OpenAlex)

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 changedWhereMeasuredEffect
use OpTop chi-square test instead of perplexity for choosing topic countLDA topic count selection · Simulated corpora (Ksim=10)proportion closer to true K: 0.84Improved
use OpTop instead of perplexity for choosing topic countLDA topic count selection · Simulated corpora (Ksim=15)proportion closer to true K: 0.90Improved
use OpTop instead of perplexity for choosing topic countLDA topic count selection · Simulated corpora (Ksim=20 to 35)proportion closer to true K: 1.00Improved
use OpTop instead of perplexity on the inaugural corpusLDA topic count selection · US Presidential Inaugural Addressesselected number of topics: 86 -> 92Mixed
add more topics beyond the optimal 92 (93 to 200) (ablation)OpTop-selected 92-topic LDA · US Presidential Inaugural Addressestopic stability (scaled TWW chi-square): scaled means approach 1.0 versus cutoff 1.414No clear effect
add uninformative least-similar topics beyond K=92 (ablation)OpTop-selected 92-topic LDA · US Presidential Inaugural Addressesincremental explanatory powerNo clear effect

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy