A Hybrid Feature Selection Method to Improve Performance of a Group of Classification Algorithms.
Mehdi NaseriparsaAmir-Massoud BidgoliTouraj VaraeePublished in: CoRR (2014)
Keyphrases
- feature selection
- classification accuracy
- classification algorithm
- computational cost
- classification method
- support vector machine
- unsupervised learning
- preprocessing
- significant improvement
- machine learning methods
- machine learning
- model selection
- nearest neighbour
- computationally efficient
- classification scheme
- similarity measure
- pattern classification
- feature set
- machine learning algorithms
- text categorization
- support vector machine svm
- discriminative features
- cross validation
- decision trees
- ensemble classifier
- computational complexity
- theoretical analysis
- text classification
- mutual information
- training samples
- feature extraction
- knn
- objective function
- support vector
- pairwise
- feature space
- select relevant features
- method for feature selection
- redundant features
- markov blanket
- classification process
- feature selection algorithms
- high dimensionality
- clustering algorithm
- multi class