A probabilistic kernel approach for solving the multi-instance learning problems with different assumptions.
Lixin ShenJianjun HeShuang QiaoPublished in: Int. J. Adv. Oper. Manag. (2013)
Keyphrases
- learning problems
- multi instance
- kernel methods
- binary classification
- semi supervised learning
- multiple kernel learning
- learning tasks
- reproducing kernel hilbert space
- multi label
- real valued
- learning algorithm
- machine learning algorithms
- supervised learning
- multiple instance
- semi supervised
- kernel function
- support vector
- unlabeled data
- machine learning
- probabilistic model
- generative model
- reinforcement learning
- text categorization
- labeled data
- unsupervised learning
- multi task
- bayesian networks
- class labels
- text classification
- multi class
- image segmentation
- graph cuts