Environment for Identification of Significant Subjects on Information Portals

摘要

A large amount of information available on the Internet makes it difficult to identify the most significant or interesting (for a given user) facts and relations between them. It especially concerns the content of information portals, e.g. provided by popular newspapers and magazines. A pilot version of the software environment for aggregating such information and extracting interesting topics and relationships between them for the user is proposed. Techniques of multi-agent approach, data mining and complex network analysis are used. Evaluation is done by analysing a set of press notes made available by the world's most popular magazines and contents provided by Polish information portals.