Chinese Document Information Processing Model Based on Random Walk Algorithm

In this paper, we conduct research on Chinese document information processing model based on random walk algorithm. Because of the complexity and also the particularity of processing Chinese information, Chinese search engine technology needs to be improved. The Chinese search engine cannot directly copy foreign technology. To study and analyze the expertise of the Chinese, we can accurately find the need in vast information base as the Chinese information. In this paper, the dictionary learning and sparse representation with random walk model are introduced into the character recognition to solve the problem of pen character and noise of the fax characters. The novel analytic framework is presented to assist the processing of the methodologies. The recognition method does not require preprocessing operations such as character binarization and thinning, only one feature and one classifier is needed, compared with the current multi-feature multi-cascade classifier fusion recognition method, proposed recognition method has characteristics of low complexity. The test on the experiment also reflects the robustness of the proposed model.