Kernel methods for software effort estimation - Effects of different kernel functions and bandwidths on estimation accuracy.
Ekrem KocaguneliTim MenziesJacky W. KeungPublished in: Empir. Softw. Eng. (2013)
Keyphrases
- kernel methods
- kernel function
- estimation accuracy
- support vector
- kernel matrices
- learning problems
- kernel matrix
- feature space
- high dimensional feature space
- support vector machine
- estimation error
- svm classifier
- support vector regression
- prediction accuracy
- input space
- learning tasks
- reproducing kernel hilbert space
- kernel parameters
- kernel trick
- kernel pca
- gaussian kernel
- gaussian kernels
- kernel learning
- similarity function
- kernel principal component analysis
- multiple kernel learning
- high dimensional
- hyperplane
- multiple kernel
- polynomial kernels
- machine learning
- positive semidefinite
- feature set
- missing data
- feature selection