Domain-specific Noisy Query Correction using Linguistic Network Community Detection

摘要: Noisy queries pose an important challenge for retrieving relevant search results. The importance query correction increases with increasing use of hand-held devices and technologies such as SMS, tweets to access information. task is further complicated domain-specific engines the amount logs may be significantly smaller than general purpose engines. In this paper, we propose community detection technique from social network analysis spelling a set noisy SMS messages. We focus on identifying questions Frequently Asked Questions (FAQ) different domains incoming queries. Experimental validation shows that proposed CD-Speller method performs better Hunspell, popular industry-strength tool.