A proposed method to evaluate and compare fault predictions across studies.
Barbara RussoPublished in: PROMISE (2014)
Keyphrases
- preprocessing
- high accuracy
- cost function
- similarity measure
- significant improvement
- experimental evaluation
- detection method
- clustering method
- synthetic data
- mutual information
- support vector machine
- computational cost
- high precision
- multiscale
- neural network
- classification method
- error rate
- evaluation method
- detection algorithm
- input data
- optical flow
- pairwise
- learning algorithm