A feature weighted penalty based dissimilarity measure for k-nearest neighbor classification with missing features.
Shounak DattaDebaleena MisraSwagatam DasPublished in: Pattern Recognit. Lett. (2016)
Keyphrases
- dissimilarity measure
- feature set
- feature vectors
- image features
- feature space
- dissimilarity representation
- invariant features
- significant features
- feature values
- missing data
- predictive features
- additional features
- redundant features
- classification accuracy
- feature extraction
- single feature
- shape recognition
- discriminative features
- feature interactions
- irrelevant features
- feature representation
- multiple features
- feature descriptors
- pattern recognition
- machine learning
- image similarity
- database
- image pairs
- training data
- selecting features
- computer vision
- neural network