Text Classification with Named-Entity Recognition and AutoPhrase
- Collaborated with a team to design and experiment with a pre-trained NER model and AutoPhrase-based model on two text datasets.
- Analyzed performance across different feature extraction methods under supervised setting.
- Achieved 96.4% accuracy on BBC News Dataset (5 classes) and 84.83% on 20 News Dataset (20 classes).
- Webpage
- Download paper here