There is Incremental Learning in the chaotic neural network's learning.
This study examined effects of the binary ratio of 1 and in patterns on Incremental Learning.
The ratio of 1 is changed from 20% to 49% and at each ratio random patterns are generated. For each ratio, 10 sets consisting of 300 patterns are made and the number of patterns that the network has succeeded in learning was averaged.
As a result, it was found that the average number of learning success increases as the ratio gets closer to 20:80 or 50:50 from the boundary ratio of 40:60.
Therefore, the correlation value of the binaly ratio of the input patterns and the nature of the ratio of 1:1 have the effect on the learning.