LEARNING MONOTONIC-CONCAVE INTERVAL CONCEPTS USING THE BACKPROPAGATION NEURAL NETWORKS

Monotonicity and concavity play important roles in human cognition, re asoning, and decision making. This paper shows that neural networks ca n learn monotonic-concave interval concepts based on real-world data. Traditionally, the training of neural networks has been based only on raw data. in cases where the training samples carry statistical fluctu ations, the products-of the training have often suffered. This paper s uggests that global knowledge about monotonicity and concavity of a pr oblem domain can be incorporated in neural network training. This pape r proposes a learning scheme for the hack-propagation layered neural n etworks in learning monotonic-concave interval concepts and provides a n example to show its application.