Synonymous Chinese transliterations retrieval from world wide web by using association words

Abstract

We present a framework for mining synonymous transliterations from a set of Web pages collected via a search engine. An integrated statistical measure is proposed to form search keywords for a search engine in order to retrieve relevant Web snippets. We employ a scheme of comparing the similarity between two transliterations to aid in identifying synonymous transliterations. Experimental results show that the average number of harvesting synonymous transliterations is about 5.04 for an input transliteration. The retrieval results could be beneficial for constructing ontology, especially, in the domain of foreign person names. © 2008 Springer-Verlag Berlin Heidelberg.