What type of study is this?

September 10, 2025Программные системы и вычислительные методыOpen Access

Two-step semantic clustering of embeddings as an alternative to LDA for infometric analysis of industry news.

View Full Paper

Ask AI

Bookmark

Authors

EKEvgenii KonnikovNational Research Institute for Veterinary Virology and Microbiology of Russia DKDarya KryzhkoPeter the Great St. Petersburg Polytechnic University

Discussion

Member takes

Implication

This approach demonstrates improved coherence and reduced perplexity in analyzing industry news, indicating the merits of geometric embedding methods.

Key Points

The proposed method shows a 5% reduction in perplexity, enhancing the interpretability of thematic modeling in news analysis.
Comparative analysis with LDA indicates a 0.15-point increase in topic coherence, highlighting improvements in thematic diversity.
The two-step semantic smoothing approach leverages embeddings and clustering techniques to form document thematic profiles effectively.
This technique's computational efficiency allows it to perform effectively in resource-constrained environments, supporting online news monitoring.

Cite This Study

Konnikov et al. (2025) studied this question.

synapsesocial.com/papers/68c1ac0154b1d3bfb60e45b5 https://doi.org/10.7256/2454-0714.2025.3.75348

View Full Paper

Ask AI

Bookmark

Implication

This approach demonstrates improved coherence and reduced perplexity in analyzing industry news, indicating the merits of geometric embedding methods.

Two-step semantic clustering of embeddings as an alternative to LDA for infometric analysis of industry news.

Discussion

Member takes

Implication

Key Points

Cite This Study

Two-step semantic clustering of embeddings as an alternative to LDA for infometric analysis of industry news.

Discussion

Member takes

Implication

Key Points

Cite This Study

Also Consider

Also Consider