A novel characterisation-based algorithm to discover new knowledge from classification datasets without use of support.
Enrique LazcorretaFederico BotellaAntonio Fernández-CaballeroPublished in: Expert Syst. Appl. (2018)
Keyphrases
- learning algorithm
- classification algorithm
- preprocessing
- computational complexity
- expectation maximization
- np hard
- cost function
- matching algorithm
- knowledge management
- worst case
- dynamic programming
- model selection
- detection algorithm
- benchmark datasets
- objective function
- synthetic datasets
- expert systems
- feature selection algorithms
- support vector machine svm
- optimal solution
- similarity measure
- computational cost
- feature selection
- genetic algorithm
- image classification
- feature subset
- classification method
- multi class classification
- extracted features
- classification rate
- training dataset
- data sets
- domain knowledge
- k means
- feature vectors
- training set
- feature space
- feature extraction
- knowledge base
- data mining