Back to Search
Start Over
Prediction Focused Topic Models via Feature Selection
- Publication Year :
- 2019
-
Abstract
- Supervised topic models are often sought to balance prediction quality and interpretability. However, when models are (inevitably) misspecified, standard approaches rarely deliver on both. We introduce a novel approach, the prediction-focused topic model, that uses the supervisory signal to retain only vocabulary terms that improve, or at least do not hinder, prediction performance. By removing terms with irrelevant signal, the topic model is able to learn task-relevant, coherent topics. We demonstrate on several data sets that compared to existing approaches, prediction-focused topic models learn much more coherent topics while maintaining competitive predictions.<br />Comment: AISTATS 2020. arXiv admin note: substantial text overlap with arXiv:1911.08551
Details
- Database :
- arXiv
- Publication Type :
- Report
- Accession number :
- edsarx.1910.05495
- Document Type :
- Working Paper