Signature-Based Chinese Character Retrieval Schemes Using DAYI Input Method

Abstract

In this paper, we propose two new Chinese character schemes using the DAYI input method. Our schemes are based on signature files, which have always been used for text retrieval. These schemes are economical in storage space use and are very efficient when dealing with large databases. Moreover, the implementation of wild-character function becomes easy and quick. The problem one of our schemes has to face is that it produces false drops. That is, some DAYI codes indicating Chinese characters many in fact come out to find no characters to them at all. We argue that this is not a big fault in real applications. On the other hand, we also propose an alternative method to retrieve Chinese characters, which efficiently solves the false drop problem but takes more space and access time. In this paper, we make an experiment for it, a comparison between traditional method and the methods we propose has been made to show the superiority of our schemes.