Query Expansion using Phonetic Confusions for Chinese Spoken Document Retrieval

摘要

This paper presents a method of query expansion based on phonetic confusions for retrieving spoken documents using text queries. This method is applied to a Chinese spoken document retrieval task. A series of experiments have been carried out for Cantonese broadcast news documents using a multi-scale syllable-based retrieval approach. Our results show an improvement from AIR (average inverse rank) of 0.481 to 0.491 when we apply query expansion based on phonetic confusions to our retrieval task.