Slavko Žitnik (2010) Intelligent data retrieval from arbitrary data sources. EngD thesis.
Search engine is a system used by everyone. When not using just web search engine, we definitely have used some other kind of it. Every larger corpus of items has its own style of search engine, just like table of contents in a book. The aim of this diploma thesis is to present a new search engine and information retrieval system respectively. We have named it iCORE - intelligent context-aware (multisource) information retrieval. This system personalizes search even more than others. It is structured in the way that can be used on any possible device to also track the user. iCORE wants to focus on the user and retrieve relevant information even before user actually tries to give an explicit input. It uses other specialized search engines or data collections. Results are presented as structured information, which are suitably shown. The user has to have options such as setting his preferences, faceted search and navigating through results. System is based on the state-of-the-art algorithms, also used in machine learning. Therefore each component can be tested using known datasets. The best way to evaluate iCORE system as a whole is just by starting using it for real purposes. During search engine developments there have been built many meta-search engines, all of which have on average returned better results. iCORE system is a lot more than just a pure meta-search engine, so it will definitely improve results and user experience.
Actions (login required)