GpLSI extends frequentist pLSI with graph total-variation denoising of left singular vectors, yielding improved topic mixture estimation on short documents and high-probability error bounds under low-p and anchor-document assumptions.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Graph Topic Modeling for Documents with Spatial or Covariate Dependencies
GpLSI extends frequentist pLSI with graph total-variation denoising of left singular vectors, yielding improved topic mixture estimation on short documents and high-probability error bounds under low-p and anchor-document assumptions.