Home Knowledge Base Latent Dirichlet Allocation

Latent Dirichlet Allocation

No mentions found

This entity hasn't been tracked yet, or Iris is still building its knowledge base.

Related Articles from SNS

Disentangling Similarity and Relatedness in Topic Models

Announce Type: replace Abstract: The recent success of large pre-trained language models (PLMs) has motivated their integration into topic modeling. However, PLM-augmented topic models differ from classical co-occurrence models such as Latent Dirichlet Allocation (LDA) not only in performance, but also in the type of semantic structure they capture. We formalize this distinction along two psycholinguistic axes: thematic relatedness (dog/bone) and taxonomic similarity (dog/wolf).

arXiv CS 8d ago

A Systematic Evaluation of Current Architectures in Wind Power Forecasting

arXiv:2606.02849v1 Announce Type: new Abstract: Interval wind speed forecasting is essential for the efficient integration of wind energy into power systems, as it accounts for the inherent uncertainty of wind resources. This study presents a systematic literature review focused on hybrid approaches to interval forecasting of wind generation, exploring the combination of deep learning, modal decomposition, and statistical methods. To guide the paper selection, Latent Dirichlet Allocation...

arXiv CS 7d ago

Topic modeling reveals thermally partitioned and taxonomically distinct microbial subcommunities across prokaryotes and phytoplankton in the Laurentian Great Lakes

Identifying discrete microbial assemblages and their environmental drivers across multiple biological fractions simultaneously remains a central challenge in aquatic microbial ecology. We applied an integrated analytical pipeline built around Latent Dirichlet Allocation (LDA) to an eight-year 16S rRNA amplicon time series from the Laurentian Great Lakes, spanning four size-fractionated biological blocks free-living prokaryotes, particle-associated prokaryotes, and small and large...

bioRxiv 2d ago

REStack: A Large-Scale Dataset of Reverse Engineering Discussions from Stack Exchange

arXiv:2606.05493v1 Announce Type: new Abstract: Reverse engineering (RE) is a critical activity in software engineering and cybersecurity, supporting tasks such as malware analysis, vulnerability discovery, legacy system maintenance, and firmware inspection. Despite its importance, there is limited empirical understanding of the challenges, topics, and knowledge gaps faced by RE practitioners in real-world settings, and no publicly available dataset has systematically captured RE discussions...

arXiv CS 5d ago