Meta search engine powered by DBpedia

• Main Authors: Chin, Kim On, Patricia Anthony, Boo, Vooi Keong, Rosalam Sarbatly
• Format: Research Report
• Language: English
• Published: Universiti Malaysia Sabah 2012
• Subjects: TA Engineering (General). Civil engineering (General)
• Online Access: https://eprints.ums.edu.my/id/eprint/22829/1/Meta%20search%20engine%20powered%20by%20DBpedia.pdf; https://eprints.ums.edu.my/id/eprint/22829/

The difficulty of information retrieval on the web is proportional to the size of information available on the web. Nowadays infonnation on the web has grown to a size at which even a good ranking algorithm could not produce a precise search result. The disambiguation that exists between several terms gives challenge on how a search engine should produce a search result since traditional search engines work based on pattern matching rather than the meaning of the term being queried. On the other hand, semantic search engines search for concept that focuses on the meaning of the input query, rather than considering a query as a group of string. However, the Web is still dominated by Web 2.0 in which infonnation and data is presented in an unstructured manner and is only fitfor human consumption. Hence, building a semantic search engine is a very challenging task and there are still a lot of improvements that needs to be done especially infield of Natural Language Processing (NLP) to achieve the desirable results. Hence, this research proposes a semantic meta search engine that utilizes the power of a traditional search engine and enriches the search result by trying to understand the meaning of the search query. This is achieved by making use of DBpedia, an RDF triple dataset that is derived from structured information on Wikipedia which gives a huge graph of how concepts are related to each other. The proposed meta search engine takes into account the design of Yippy, a meta search engine that has of the ability to cluster search result into a group of topics that may be relevant to the query. This semantic meta search engine also uses Google to generate the search results of a given query, explores the meaning of the query using DBPedia andfinally displays the results in cluster similar to the technique used by Yippy. It also provides broad suggestions based on concepts that are related the query. The experimental evaluation showed that by using DBpedia dataset, the non-semantic search results can be clustered, rearranged and enriched with a certain degree of semantic. The proposed search engine is also able to provide suggestions on concepts that are similar to the query.