Submodular feature selection for high-dimensional acoustic score spaces.
Yuzong LiuKai WeiKatrin KirchhoffYisong SongJeff A. BilmesPublished in: ICASSP (2013)
Keyphrases
- high dimensional
- feature selection
- high dimensionality
- feature space
- dimensionality reduction
- microarray data
- high dimension
- small sample
- gene expression data
- low dimensional
- pointwise mutual information
- dimension reduction
- text categorization
- greedy algorithm
- feature set
- discrete data
- similarity search
- sparse data
- source localization
- high dimensional data
- energy minimization
- low dimensional spaces
- data points
- feature extraction
- feature selection algorithms
- variable selection
- input space
- classification accuracy
- mutual information
- support vector machine
- machine learning
- high order
- kernel function
- feature subset
- objective function
- information gain
- acoustic signal
- text classification
- discriminative features
- unsupervised feature selection
- multi dimensional
- speech segments
- method for feature selection
- underwater acoustic
- training data
- knn
- scoring function
- nearest neighbor
- noisy data
- manifold learning