A COBOL-like neural network simulation language based on layer macro definition

Most of neural network simulation programs have been implemented with the conventional computer languages. It is difficult to implement simulation programs because the languages have no data structures and statements suitable for expressing neural network structures and controlling neural behaviors. This paper describes a neural network simulation language, specifically designed for the development and modeling of neural network applications. It has a layer macro definition concept to express the arbitrary network structures and dynamic neural behaviors easily. It consists of two divisions: the first division describes the network structures, and the second division controls the network behaviors.