Research of simulating recognition of Chinese characters by Bidirectional SOFM
利用双向SOFM网络模拟汉字认知过程的研究

Word superiority effect and word inferiority effect are invovled in recognition of Chinese characters. The former one means that the recognition of holistie character unit is better than that of radical unit and the latter contrarily. In this paper, bidirectional SOFM was proposed based on Self-Organization Feature Map Mgorithm, accomplishing clustering of Chinese characters and their radical unit, establishing the connection between the two nets. Through bidirectional test, diversity constructional Chinese characters have their own effects, and SOFM is used in a new perspective. The result is helpful to Chinese characters teaching.