What Number of Features is Optimal: A New Method Based on Approximation Function for Stance Detection Task.
Sergey V. VychegzhaninElena V. RazovaEvgeniy V. KotelnikovPublished in: ICICM (2019)
Keyphrases
- detection method
- classification method
- detection algorithm
- dynamic programming
- high accuracy
- closed form
- cost function
- similarity measure
- feature vectors
- computational complexity
- computational cost
- classification accuracy
- feature set
- linear approximation
- image regions
- single parameter
- linear functions
- support vector machine classifier
- approximation error
- exhaustive search
- roc curve
- neural network
- basis functions
- segmentation method
- false positives
- model selection
- image features
- pairwise
- feature extraction