Real-time web mining application to support decision-making process
Jan Hovad, Martin Lněnička, Jitka Komárková · 2015
This paper describes a solution for the improvement and support of decision-making processes in a reasonable time. It proposes and implements a suitable procedure based on appropriate methods and technologies to create a Web content mining application focused on the trend analysis. The following text describes used architecture, methods and data structures in the context of Web mining and the decision-making approaches. The resulted code runs as a timed script that handles different analyses over the selected Web sites - e.g. the most covered topics in the mass media servers, contexts of the top words based on the distance, saving outputs into the graph, etc. Authors are able to compare options and make decisions very fast, without reading the full contents of the sources. Object-oriented approach is used for the code writing so the code can be easily extended to handle number of different tasks (financial market decisions, crisis management, customer segmentation, social media monitoring, etc.). Real-time analysis of data from different Internet sources and a graphical visualization of results belong to the most important contributions of the proposed application.