Approach and Visualization Strategies of Big Data Using Semantic Graphs

Abstract

The proposal made here is to design and develop a document similarity measurement tool, contemplating the semantic analysis and allowing the visualization of its relationship scheme, to facilitate the final user the selection from a scientific documentary base. For this, various existing possibilities will be analyzed and compared in terms of the design of the internal support data structures of the search engine (vector model, probabilistic model, variants of the ranking model by popularity, or even variants based on application of meta- search engines that process the results delivered by two or more pre-existing engines), while new variants will be proposed that can better fit the specific context of this research. In the area of visualization of relations between documents, work will be done on the design of a graphical interface model that represents the recovered documents as visually manageable objects, also highlighting visually the similarity relations between them.