Reading Tea Leaves: How Humans Interpret Topic Models
Explore this paper's citation graph
Summary
New quantitative methods for measuring semantic meaning in inferred topics are presented, showing that they capture aspects of the model that are undetected by previous measures of model quality based on held-out likelihood.
- Type
- article
- Published
- 2009-12-07
- Cited by
- 2,563
- References
- 30
- OpenAlex
- https://openalex.org/W2159426623
- Semantic Scholar
- https://api.semanticscholar.org/CorpusID:215812433
Keywords
Topic model, Computer science, Latent semantic analysis, Natural language processing, Artificial intelligence
References
- Probabilistic Latent Semantic Analysis
- Cheap and Fast – But is it Good? Evaluating Non-Expert Annotations for Natural Language Tasks
- A Solution to Plato's Problem: The Latent Semantic Analysis Theory of Acquisition, Induction, and Representation of Knowledge.
- Nouns in WordNet: A Lexical Inheritance System
- LDA-based document models for ad-hoc retrieval
- Maximum likelihood from incomplete data via the EM - algorithm plus discussions on the paper
- Mining business topics in source code using latent dirichlet allocation
- ImageNet: A large-scale hierarchical image database
- Correlated Topic Models
- Automatic labeling of multinomial topic models
- Computational Methods for Intelligent Information Access
- Organizing the OCA: learning faceted subjects from a library of digital books
- NLTK: The Natural Language Toolkit
- Mixed-membership models of scientific publications
- A Joint Model of Text and Aspect Ratings for Sentiment Summarization
- Collapsed Variational Inference for HDP
- Studying the History of Ideas Using Topic Models
- Semantic Similarity Based on Corpus Statistics and Lexical Taxonomy
- NLTK: The Natural Language Toolkit
- Evaluation methods for topic models
Cited by
- Von Mises-Fisher Clustering Models
- Summarization of Corporate Risk Factor Disclosure through Topic Modeling
- Comparing Taxonomies for Organising Collections of Documents
- Universal Schema for Knowledge Representation from Text and Structured Data
- Leveraging Domain Knowledge in Multitask Bayesian Network Structure Learning
- Enriching iTunes App Store Categories via Topic Modeling
- Visualizing Topic Models
- SurfShop: combing a product ontology with topic model results for online window-shopping.
- Modeling Mortality Rates In The WikiLeaks Afghanistan War Logs
- Topic Model Diagnostics: Assessing Domain Relevance via Topical Alignment
- Unsupervised Part-of-Speech Tagging in Noisy and Esoteric Domains With a Syntactic-Semantic Bayesian HMM
- 3D Robotic Sensing of People: Human Perception, Representation and Activity Recognition
- Best Topic Word Selection for Topic Labelling
- Software Framework for Topic Modelling with Large Corpora
- Cooperative Semantic Information Processing for Literature-Based Biomedical Knowledge Discovery
- Bayesian Checking for Topic Models
- Thread labeling for news event
- Graph-Sparse LDA: A Topic Model with Structured Sparsity
- Convex Approaches to Text Summarization
- Examining the Impact of Keyword Ambiguity on Search Advertising Performance: A Topic Model Approach
Related papers
- GloVe: Global Vectors for Word Representation
- Machine Reading Tea Leaves: Automatically Evaluating Topic Coherence and Topic Model Quality
- Probabilistic topic models
- Online Learning for Latent Dirichlet Allocation
- Hierarchical Dirichlet Processes
- Automatic Evaluation of Topic Coherence
- Indexing by Latent Semantic Analysis
- Evaluation methods for topic models
- The author-topic model for authors and documents
- Hierarchical Topic Models and the Nested Chinese Restaurant Process