Sailing the Web with Captain Nemo: A personalized metasearch engine

Personalization on the Web is an issue that has gained a lot of interest lately. Web sites have already started providing services such as preferences for the interface, the lay-out and the functionality of the applications. Personalization services have also been intro-duced in Web search and metasearch engines, i.e. tools that retrieve Web pages relevant to keywords given by the users. However, those services deal mostly with the presenta-tion style and ignore issues like the retrieval model, the ranking algorithm and topic pref-erences. In this paper, we present Cap-tain Nemo, a fully-functionable metasearch engine that exploits personal user search spaces. Users can define their personal re-trieval model and presentation style. They can also define topics of interest. Captain Nemo exploits several popular Web search engines to retrieve Web pages relevant to key-words given by the users. The resulting pages are presented according to the defined pre-sentation style and retrieval model. For every page, Captain Nemo can recommend a rele-vant topic of interest to classify the page, ex-ploiting nearest-neighbour classification tech-niques. 1