Parallel feature selection using positive approximation based on MapReduce.
Qing HeXiaohu ChengFuzhen ZhuangZhongzhi ShiPublished in: FSKD (2014)
Keyphrases
- feature selection
- parallel processing
- parallel programming
- high performance data mining
- parallel computing
- distributed processing
- text categorization
- error bounds
- positive and negative
- multi class
- parallel implementation
- text classification
- machine learning
- data parallelism
- feature set
- mutual information
- information gain
- feature selection algorithms
- classification accuracy
- irrelevant features
- massively parallel
- unsupervised feature selection
- parallel computation
- support vector
- data partitioning
- closed form
- approximation algorithms
- support vector machine
- approximation methods
- distributed memory
- feature ranking
- discriminative features
- feature space
- shared memory