Using Weighted Nearest Neighbor to Benefit from Unlabeled Data.
Kurt DriessensPeter ReutemannBernhard PfahringerClaire LeschiPublished in: PAKDD (2006)
Keyphrases
- unlabeled data
- nearest neighbor
- data points
- labeled data
- training set
- semi supervised learning
- semi supervised
- co training
- active learning
- supervised learning
- semi supervised classification
- training data
- decision boundary
- text classification
- k nearest neighbor
- knn
- high dimensional
- high dimensional data
- labeled and unlabeled data
- number of labeled examples
- small set of labeled
- class labels
- text categorization
- neural network
- distance function
- domain adaptation
- labeled examples
- prior knowledge
- labeled training data
- label propagation
- positive examples
- learning algorithm
- index structure
- training examples
- feature space
- transfer learning
- instance selection
- unsupervised learning
- multi view
- information extraction
- labeled data for training
- machine learning algorithms
- knowledge discovery
- reinforcement learning
- select relevant features