Recognition of handwritten Chinese characters based on association rules
Jun Wang · Jisuanji gongcheng · 2004
A new improved method, named weighted AFDC, which can overcome the drawbacks of traditional fuzzy methods, was presented to enhance the robustness and generalization of the recogintion algorithm. Some changes about association rules in DM (Data Mining) were made to adapt to generalizing the warehouse of standard samples so as to attain special attributes, division headers, and achieve the process of evolution dynamically. Therefore with a database of division headers ordered by support degree, the Chinese character recognition system will become more adapted, more flexible, more extendible and more intelligent while a more reasonable expression of Chinese character was created.