BERTopic: Neural topic modeling with a class-based TF-IDF procedure

Abstract

Topic models can be useful tools to discover latent topics in collections of documents. Recent studies have shown the feasibility of approach topic modeling as a clustering task. We present BERTopic, a topic model that extends this process by extracting coherent topic representation through the development of a class-based variation of TF-IDF. More specifically, BERTopic generates document embedding with pre-trained transformer-based language models, clusters these embeddings, and finally, generates topic representations with the class-based TF-IDF procedure. BERTopic generates coherent topics and remains competitive across a variety of benchmarks involving classical models and those that follow the more recent clustering approach of topic modeling.

arXiv commentBERTopic has a python implementation, see https://github.com/MaartenGr/BERTopic

Cited by in corpus (50)

• Understanding Environmental Posts: Sentiment and Emotion Analysis of Social Media Data
• Using Twitter Data to Understand Public Perceptions of Approved versus Off-label Use for COVID-19-related Medications
• Drivers of social influence in the Twitter migration to Mastodon
• The Systemic Impact of Deplatforming on Social Media
• pyBibX -- A Python Library for Bibliometric and Scientometric Analysis Powered with Artificial Intelligence Tools
• Beyond Text-to-Text: An Overview of Multimodal and Generative Artificial Intelligence for Education Using Topic Modeling
• CrisisTransformers: Pre-trained language models and sentence encoders for crisis-related social media texts
• Advanced Deep Learning and Large Language Models: Comprehensive Insights for Cancer Detection
• PromptMTopic: Unsupervised Multimodal Topic Modeling of Memes using Large Language Models
• Multilingual transformer and BERTopic for short text topic modeling: The case of Serbian
• The Elements of Visual Art Recommendation: Learning Latent Semantic Representations of Paintings
• Angler: Helping Machine Translation Practitioners Prioritize Model Improvements
• DeTiME: Diffusion-Enhanced Topic Modeling using Encoder-decoder based LLM
• Word embeddings for topic modeling: an application to the estimation of the economic policy uncertainty index
• Weakly-Supervised Scientific Document Classification via Retrieval-Augmented Multi-Stage Training
• Public Attitudes Toward ChatGPT on Twitter: Sentiments, Topics, and Occupations
• Large Language Models for Scholarly Ontology Generation: An Extensive Analysis in the Engineering Field
• Wayfinding through the AI wilderness: Mapping rhetorics of ChatGPT prompt writing on X (formerly Twitter) to promote critical AI literacies
• SuperNOVA: Design Strategies and Opportunities for Interactive Visualization in Computational Notebooks
• Exploring the evolution of research topics during the COVID-19 pandemic
• Multilingual Previously Fact-Checked Claim Retrieval
• Big Tech influence over AI research revisited: memetic analysis of attribution of ideas to affiliation
• Unveiling Novelty Evolution in the field of Library and Information Science in China
• vONTSS: vMF based semi-supervised neural topic modeling with optimal transport
• Emergency Department Decision Support using Clinical Pseudo-notes
• Enhancing Procedural Writing Through Personalized Example Retrieval: A Case Study on Cooking Recipes
• A Hybrid Intelligence Method for Argument Mining
• Recommender systems and reinforcement learning for human-building interaction and context-aware support: A text mining-driven review of scientific literature
• Exploring Public Attention in the Circular Economy through Topic Modelling with Twin Hyperparameter Optimisation
• Assessing the Potential of Generative Agents in Crowdsourced Fact-Checking
• TickIt: Leveraging Large Language Models for Automated Ticket Escalation
• Coconut Libtool: Bridging Textual Analysis Gaps for Non-Programmers
• Topic Shifts as a Proxy for Assessing Politicization in Social Media
• A Twitter narrative of the COVID-19 pandemic in Australia
• Investigating the heterogenous effects of a massive content moderation intervention via Difference-in-Differences
• Science and engineering for what? A large-scale analysis of students' projects in science fairs
• Examining European Press Coverage of the Covid-19 No-Vax Movement: An NLP Framework
• Analysing State-Backed Propaganda Websites: a New Dataset and Linguistic Study
• Measuring Online Emotional Reactions to Events
• pytopicgram: A library for data extraction and topic modeling from Telegram channels
• HINTs: Sensemaking on large collections of documents with Hypergraph visualization and INTelligent agents
• Examining the Representation of Youth in the US Policy Documents through the Lens of Research
• Automatic Generation of Web Censorship Probe Lists
• SeLeRoSa: Sentence-Level Romanian Satire Detection Dataset
• Zero-Truncated Poisson Regression for Sparse Multiway Count Data Corrupted by False Zeros
• Improving Customer Service with Automatic Topic Detection in User Emails
• Multi-Faceted Question Complexity Estimation Targeting Topic Domain-Specificity
• MALTopic: Multi-Agent LLM Topic Modeling Framework
• Image-guided topic modeling for interpretable privacy classification
• Social Media and Artificial Intelligence for Sustainable Cities and Societies: A Water Quality Analysis Use-case