Computer simulation of the cognition of Chinese characters
Xiao-qian Sun · Caai Transactions on Intelligent Systems · 2008
Research on the cognition of Chinese characters is a hotspot in both cognitive science and computer science,and is an especially lively field among those investigating artificial intelligence.In spite of this,research using computer simulations to analyze Chinese character cognition remains relatively backward,and its value in cognition studies has not been comparable with experimental research on behavior.In this paper,starting from the viewpoint of cognition science,a representative database of Chinese characters was set up,a cognitive model constructed,and training and testing modes determined.Computer simulations were made of the clustering and splitting of Chinese characters in the course of cognition,so that cognition rules for the perception of Chinese characters may be better understood.The model was based on a multi-layer self-organizing neural network.This training and testing method ensured that we knew how the Chinese characters were clustered and split during analysis so that the recognition of unknown words could be achieved.The research outcome suggests cognition rules for recognizing Chinese characters,implying that the proposed model does simulate the cognition process for Chinese characters.