Using a Penalty-based Loss Re-estimation Method to Improve Implicit Discourse Relation Classification.
Xiao LiYu HongHuibin RuanZhen HuangPublished in: COLING (2020)
Keyphrases
- classification method
- classification accuracy
- high accuracy
- classification process
- computational cost
- preprocessing
- support vector machine svm
- cross validation
- classification scheme
- pairwise
- model selection
- image classification
- classification procedure
- pattern classification
- benchmark data sets
- unsupervised learning
- estimation algorithm
- accuracy rate
- data sets
- support vector machine
- dynamic programming
- significant improvement
- objective function
- similarity measure
- machine learning
- em algorithm
- training samples
- decision rules
- classification algorithm
- feature vectors
- image sequences
- image segmentation
- classification rate
- feature selection
- genetic algorithm