Handwritten Chinese Character Recognition with Variable Precision Rough Set Approach

To study the recognition of off-line handwritten Chinese character, a kind of decision-making information system is constructed. The character feature is considered as condition attribute, and the true value of character samples is regard as decision attribute. A heuristic algorithm of the attribute reduction based on variable precision rough set is presented for reducing the redundancy features. The attribute importance is defined based on β approximation dependence, it is taken as heuristic information in that algorithm. In the case of no rule can be fully matched in the process of recognition, to improve the identifiability and correct classification rate of handwritten Chinese character, a kind of rules fusion method based on weighted rules confidence is proposed. The experiment results show that the method is feasible and effective.