A neural network model for cognition of the order relation among sequential input patterns : A basic idea on forecast and abstraction(Application,Scientific Computation and Numerical Analysis,<Special Issue>Joint Symposium of JSIAM Activity Groups 2007)

Abstract

We propose a neural network model which commits the order relation among sequential input patterns to memory, using associative memory for sequential patterns. The network is constructed by association of Leaky Integrate and Fire neuron models. The model may give the basic concept to some functions of the brain, especially the functions of forecast and abstraction. We implement our model to computer and investigate the numerical results in order to consider what the model gives to the concept of forecast and abstraction.