The Perceptron Algorithm Versus Winnow: Linear Versus Logarithmic Mistake Bounds when Few Input Variables are Relevant (Technical Note).
Jyrki KivinenManfred K. WarmuthPeter AuerPublished in: Artif. Intell. (1997)
Keyphrases
- perceptron algorithm
- mistake bound
- input variables
- multi class
- learning algorithm
- winnow algorithm
- hyperplane
- weight vector
- linear threshold
- worst case
- variable selection
- concept classes
- artificial neural networks
- linear program
- pac learning
- kernel classifiers
- membership functions
- fuzzy rules
- neural network model
- genetic algorithm
- concept class