A prediction method for job runtimes on shared processors: Survey, statistical analysis and new avenues.
Menno DobberRobert D. van der MeiGer KoolePublished in: Perform. Evaluation (2007)
Keyphrases
- statistical analysis
- significant improvement
- cost function
- synthetic data
- computational cost
- pairwise
- feature set
- high accuracy
- neural network
- parallel processing
- detection method
- clustering method
- classification accuracy
- experimental evaluation
- preprocessing
- similarity measure
- support vector machine
- prior knowledge
- feature space
- computationally efficient
- support vector machine svm
- detection algorithm
- objective function
- missing data
- classification method
- data sets
- correlation coefficient
- prediction error