Today, neural network attracts a great deal of attention as anti-Neumann-computer which can work simultaneously.
Neural network has some types, in association memory model, which is one of them, can raise its remembering ability by putting outputs back to itself. The model is studied to increase its ability.
Three years ago, Mr. Tachi had proposed the random-layer model mentioned above, and measured its ability.
In this paper, using this random-layer model, we compare remembering ability of the random pattern and the letter pattern, and measure remembering ability when the learning pattern increased.
As a result, we confirm that if cycles of learning pattern is large, remembering ability will improve, and this model can correct errors about 25%, and complete remembering needs the random pattern.