Proj CDeepFuzz Paper Reading: Testing and validating machine learning classifiers by metamorphic testing
Abstract
本文:Metamorphic Testing for machine learning classifiers
Method: cross-validation, metamorphic testing
- MR-0: Consistence with affine transformation
- MR-1.1: Permutation of class labels
- MR-1.2: Permutation of the attribute.
- MR-2.1: Addition of uninformative attributes.
- MR-2.2: Addition of informative attributes
- MR-3.1: Consistence with re-prediction.
- MR-3.2: Additional training sample.
- MR-4.1: Addition of classes by duplicating samples
- MR-4.2: Addition of classes by re-labeling samples
- MR-5.1: Removal of classes.
- MR-5.2: Removal of samples
实验:
- a case study on a real-world machine learning application framework