Are Your Explanations Reliable? Investigating the Stability of LIME in Explaining Textual Classification Models via Adversarial Perturbation.
Christopher BurgerLingwei ChenThai LePublished in: CoRR (2023)
Keyphrases
- classification models
- generating explanations
- training data
- feature selection
- decision trees
- visual representations
- software quality classification
- models built
- attribute selection
- learning models
- data sets
- feature set
- database
- benchmark datasets
- imbalanced data
- genetic programming
- decision tree algorithm
- semi supervised
- machine learning