Rank aggregation using ant colony approach for metasearch
Autor: | Manpreet Singh, Sukhwinder Singh Dhillon, Gurpreet Singh Josan, Parneet Kaur |
---|---|
Rok vydání: | 2017 |
Předmět: |
Information retrieval
Web search query Ant colony optimization algorithms InformationSystems_INFORMATIONSTORAGEANDRETRIEVAL 05 social sciences Rank (computer programming) Search aggregator 050301 education computer.software_genre 01 natural sciences Theoretical Computer Science Ranking (information retrieval) 010104 statistics & probability Search engine Geometry and Topology Data mining 0101 mathematics Metasearch engine 0503 education Metaheuristic computer Software Mathematics |
Zdroj: | Soft Computing. 22:4477-4492 |
ISSN: | 1433-7479 1432-7643 |
DOI: | 10.1007/s00500-017-2723-3 |
Popis: | Metasearch engines provide a plethora of information to the user through World Wide Web. They are the prominent sources of query-based search and centralized human–world interactions. Metasearch engine shows a list of Web sites to a particular query as per the rank assigned to a web link. The effectiveness of metasearch engine is also examined on the basis of ranks assigned to Web sites for a particular query. Assigning top rank to a web link with most relevant information pertaining to a query by the search engine is formulated as research problem. Here, we have formulated the rank aggregation optimization problem by using metaheuristic approach. Search engines are facing widely two problems such as biasing of search solutions and giving irrelevant rank to similar kind of documents. Both these problems can be overcome by applying an effective rank aggregation technique for combining the search results from various search engines. This paper presents a metaheuristic approach to optimize Spearman’s footrule and Kendall-tau distance measures which are used to compare ranking methods. The performance of proposed ant colony-based strategy is compared with GA technique and is validated through experimental results for real-world queries. Likewise, Precision, Recall and F-Measure-based performance metrics are employed to test the effectiveness of various metasearch engines. |
Databáze: | OpenAIRE |
Externí odkaz: |