Using Cost-Sensitive Ranking Loss to Improve Distant Supervised Relation Extraction.
Daojian ZengJunxin ZengYuan DaiPublished in: CCL (2017)
Keyphrases
- cost sensitive
- relation extraction
- multi class
- weakly supervised
- cost sensitive learning
- misclassification costs
- automatic extraction
- cost sensitive classification
- binary classification
- information extraction
- active learning
- naive bayes
- question answering
- image classification
- class distribution
- semantic relations
- domain specific
- general purpose
- feature vectors
- data sets
- named entity recognition
- semi supervised
- support vector machine
- knowledge discovery
- object recognition
- similarity measure
- decision trees