Abstract: The developed technology consists of three main stages: collection of information from a search engine; sifting irrelevant information by the pre-selected features; ranking the obtained results by relevance to a user's request. The ranking model is built with the usage of inductive1 GMDH algorithms. The article describes the effectiveness investigation of the developed technology improving the search relevance of target information on the Internet compared with the Google search engines. When studying, three experiments were conducted with one search request chosen for each experiment. The search results for every request obtained from Google were subsequently processed with the developed technology. The first 100 sites from Google SERP were analyzed to compare the relevance level of Google search and that provided with our technology. Outcomes of the experiments are given in the form of circle diagrams showing the percentage of different types of sites in the search results before and after processing it using the proposed technology. The research demonstrates higher effectiveness of the proposed technology compared to Google search: the developed technology allows achieving the search relevance at the level of 80%. Application of this technology will enable more convenient and relevant search of target information on the Internet.

Keywords: Information search, target information, search engine, search relevance, inductive modeling.

ACM Classification Keywords: H.3.5 Information Search and Retrieval - Information Filtering; H.3.5 Online Information Services – Web-based Services

Link:

ENHANCED TECHNOLOGY OF EFFICIENT INTERNET RETRIEVAL FOR RELEVANT INFORMATION USING INDUCTIVE PROCESSING OF SEARCH RESULTS

Vyacheslav Zosimov, Volodymyr Stepashko, Oleksandra Bulgakova

http://www.foibg.com/ibs_isc/ibs-26/ibs-26-p07.pdf