The role of hyper-parameters in relational topic models: Prediction capabilities vs topic quality.
Silvia TerragniAntonio CandelieriElisabetta FersiniPublished in: Inf. Sci. (2023)
Keyphrases
- topic models
- hyperparameters
- latent dirichlet allocation
- topic modeling
- variational bayes
- latent variables
- latent topics
- model selection
- generative model
- probabilistic model
- cross validation
- text documents
- bayesian inference
- bayesian framework
- text mining
- co occurrence
- random sampling
- support vector
- maximum a posteriori
- text corpora
- prior information
- sample size
- gaussian process
- posterior distribution
- latent topic models
- maximum likelihood
- incomplete data
- parameter space
- closed form
- missing values
- em algorithm
- object recognition
- probabilistic topic models
- data mining
- noise level
- incremental learning
- multiscale
- machine learning