Using the revised EM algorithm to remove noisy data for improving the one-against-the-rest method in binary text classification.
Hyoungdong HanYoungjoong KoJungyun SeoPublished in: Inf. Process. Manag. (2007)
Keyphrases
- em algorithm
- noisy data
- text classification
- expectation maximization
- maximum likelihood
- generative model
- parameter estimation
- maximum likelihood estimation
- likelihood function
- probabilistic model
- mixture model
- expectation maximisation
- mixture modeling
- gaussian mixture model
- missing data
- input data
- similarity measure
- covariance matrix
- log likelihood
- mixture distribution
- feature selection
- machine learning