Highly over-parameterized classifiers generalize since bad solutions are rare.
Julius MartinetzThomas MartinetzPublished in: CoRR (2022)
Keyphrases
- training data
- optimal solution
- decision trees
- support vector
- high dimensional
- solution quality
- test set
- classification accuracy
- solution space
- classification algorithm
- svm classifier
- class labels
- machine learning algorithms
- data sets
- linear classifiers
- learning algorithm
- training examples
- classification systems
- benchmark problems
- object detection
- feature space
- data mining