For Full-Text PDF, please login, if you are a member of IEICE,|
or go to Pay Per View on menu list, if you are a nonmember of IEICE.
Improving Search Performance: A Lesson Learned from Evaluating Search Engines Using Thai Queries
Shisanu TONGCHIM Virach SORNLERTLAMVANICH Hitoshi ISAHARA
IEICE TRANSACTIONS on Information and Systems
Publication Date: 2007/10/01
Online ISSN: 1745-1361
Print ISSN: 0916-8532
Type of Manuscript: Special Section PAPER (Special Section on Knowledge, Information and Creativity Support System)
search engine evaluation, metasearch,
Full Text: PDF>>
This study initiates a systematic evaluation of web search engine performance using queries written in Thai. Statistical testing indicates that there are some significant differences in the performance of search engines. In addition to compare the search performance, an analysis of the returned results is carried out. The analysis of the returned results shows that the majority of returned results are unique to a particular search engine and each system provides quite different results. This encourages the use of metasearch techniques to combine the search results in order to improve the performance and reliability in finding relevant documents. We examine several metasearch models based on the Borda count and Condorcet voting schemes. We also propose the use of Evolutionary Programming (EP) to optimize weight vectors used by the voting algorithms. The results show that the use of metasearch approaches produces superior performance compared to any single search engine on Thai queries.