An adaptive model for phonetic string search

Abstract

Phonetic string search of written text is an important topic in Information Retrieval (IR). A major difficulty is the inconsistencies between relevance judgements, which makes it possible for a successful method to fail with a new dataset. This paper discusses an adaptive model based on the novel syllable alignment pattern searching algorithm. Experimental results show that it is convenient and effective to be trained for different datasets. © Springer-Verlag Berlin Heidelberg 2005.