A Variance Minimization Criterion to Feature Selection Using Laplacian Regularization.
Xiaofei HeMing JiChiyuan ZhangHujun BaoPublished in: IEEE Trans. Pattern Anal. Mach. Intell. (2011)
Keyphrases
- feature selection
- selection criterion
- regularization term
- semi supervised feature selection
- half quadratic
- minimization problems
- mutual information
- objective function
- class separability
- text categorization
- feature selection algorithms
- machine learning
- graph laplacian
- regularization parameter
- selecting relevant features
- multi class
- image restoration and reconstruction
- support vector
- classification accuracy
- edge preserving
- text classification
- model selection
- feature extraction
- supervised feature selection
- tikhonov regularization
- norm minimization
- variance reduction
- irrelevant features
- unsupervised learning
- covariance matrix
- dimensionality reduction
- feature set
- sparse coding
- support vector machine
- loss function
- multi task
- unsupervised feature selection
- noise variance
- selected features
- reproducing kernel hilbert space
- edge detection
- intra class
- information gain
- feature space
- feature subset