CARISMA: a context-sensitive approach to race-condition sample-instance selection for multithreaded applications.
Ke ZhaiBoni XuW. K. ChanT. H. TsePublished in: ISSTA (2012)
Keyphrases
- context sensitive
- instance selection
- feature and instance selection
- nearest neighbor
- text classification
- data reduction
- multiple instance learning
- semi supervised learning
- multi class
- supervised learning
- knowledge discovery and data mining
- natural language
- language model
- information retrieval
- classification accuracy
- evaluation criteria
- similarity measure
- knn
- feature vectors
- high dimensional
- training set