Authority-Based Ranking in Propaganda Summary Generation Considering Values

Abstract

The era of big data provides a favorable condition for the study of news report and public opinions. By using keyword search, language processing, data mining and information techniques, websites, blogs, forums can be further analyzed for public opinion monitor and precisely guidance. Thus, in this paper, we propose a novel ranking methodology for object or propaganda summary generation, i.e., ValueRank, which incorporates tuples values in authority flow and facilitating the approximation of importance for arbitrary databases. According to ObjectRank and other PageRank fashioned techniques that consider solely authority flow through relationships and therefore their effectiveness is limited on bibliographic databases. Experimental evaluation on DBLP and Northwind databases verified the effectiveness and efficiency of our proposed approach.