Identifying the best data-driven feature selection method for boosting reproducibility in classification tasks.
Nicolas GeorgesIslem MhiriIslem RekikAlzheimer's Disease Neuroimaging InitiativePublished in: Pattern Recognit. (2020)
Keyphrases
- feature selection
- data driven
- unsupervised learning
- support vector machine
- high accuracy
- clustering method
- cost function
- feature set
- mutual information
- high precision
- detection method
- model selection
- classification accuracy
- dynamic programming
- preprocessing
- edge detection
- neural network
- multi class
- significant improvement
- segmentation method
- cross validation
- classification method
- objective function
- optimization method
- information gain
- informative features