Exploiting Association Words to Retrieve Synonymous Transliterations from Web Snippets

Abstract

There is no translation standard across the regions such as Taiwan, Hong Kong and China where Chinese language is used. As a result, a foreign proper noun is often translated to different Chinese words which lead to the incomplete search problem if only one of the words is used as the query keyword to a search engine. In this paper, we present a framework to retrieve synonymous transliterations as many as possible from the Web for an input Chinese transliteration. The research results could be applied to query expansion so as to alleviate the incomplete search problem. There are two major phases in the framework. The first is to develop an effective method to collect relevant Web snippets which may contain synonymous transliterations. The second is to extract synonymous transliterations from the set of relevant Web snippets. Experimental results show that the proposed framework is feasible and effective. Moreover, most of extracted synonymous transliterations, compared with other noise terms, have a higher rank of similarity to the input transliteration.