To Avoid the Pitfall of Missing Labels in Feature Selection: A Generative Model Gives the Answer.
Yuanyuan XuJun WangJinmao WeiPublished in: AAAI (2020)
Keyphrases
- generative model
- feature selection
- probabilistic model
- bayesian framework
- em algorithm
- missing data
- conditional random fields
- text categorization
- discriminative learning
- training data
- incomplete data
- pairwise
- unsupervised learning
- object categories
- class labels
- topic models
- semi supervised
- feature space
- machine learning
- feature set
- classification accuracy
- prior knowledge
- discriminative models
- text classification
- latent dirichlet allocation
- multi label
- posterior probability
- training set
- fisher kernel
- feature representations
- support vector
- feature subset
- knn
- feature extraction
- expectation maximization
- multi task
- image features
- training examples
- maximum likelihood