E-VarM: Enhanced Variational Word Masks to Improve the Interpretability of Text Classification Models.
Ling GeChunming HuGuanghui MaJunshuang WuJunfan ChenJihong LiuHong ZhangWenyi QinRichong ZhangPublished in: COLING (2022)
Keyphrases
- classification models
- feature selection
- training data
- decision trees
- keywords
- information retrieval
- feature set
- co occurrence
- text documents
- english words
- models built
- software quality classification
- text input
- neural network
- related words
- word pairs
- printed text
- lexical features
- english text
- text mining
- sentence level
- feature ranking
- noun phrases
- n gram
- k nearest neighbor
- search space