Interactive Semantic Featuring for Text Classification
- Patrice Simard ,
- Max Chickering ,
- Jina Suh
Proceedings of the 2016 ICML Workshop on Human Interpretability in Machine Learning (WHI 2016) |
In text classification, dictionaries can be used to define human-comprehensible features. We propose an improvement to dictionary features called smoothed dictionary features. These features recognize document contexts instead of ngrams. We describe a principled methodology to solicit dictionary features from a teacher, and present results showing that models built using these human-comprehensible features are competitive with models trained with Bag of Words features.