STRIP - a strip-based neural-network growth algorithm for learning multiple-valued functions.
Alioune NgomIvan StojmenovicVeljko MilutinovicPublished in: IEEE Trans. Neural Networks (2001)
Keyphrases
- learning algorithm
- neural network
- dynamic programming
- high accuracy
- learning phase
- preprocessing
- learning rules
- learning speed
- np hard
- detection algorithm
- computational cost
- search space
- incremental learning
- optimal solution
- simulated annealing
- optimization algorithm
- matching algorithm
- recognition algorithm
- learning process
- objective function
- convergence rate
- learning scheme
- weight update
- supervised learning
- probabilistic model
- similarity measure
- computational complexity
- significant improvement
- k means
- image segmentation
- auto associative
- feedforward neural networks
- feed forward neural networks
- lower bound
- artificial neural networks
- evolutionary algorithm
- active learning
- neural network model
- cost function
- experimental evaluation
- theoretical analysis
- online learning
- support vector machine