Homonyms's Complex Networks to Semantic Analysis Textual

Abstract

Introduction: Study centres in natural language processing already spread and the study have several applications. Relate with this research area, it is common the use technic for manipulation a text. These technic is be able to determine the word morphology and the word syntax. There are tools that do this work, however adding engines for semantic identification of the words is essential for increase the automatic understanding the used language. Objective: On the basis of that, This paper present the process of using complex networks as a comparative database to determine by context the meaning of words that express different positions. Moreover, they are classified as same morphology and syntax , as with some homonyms. Methodology: Through of a experimental methodology, the model proposed it is based in consolidate researches in Natural Language Processing for building a Complex Network that receives as vertices the words of a certain text and establishes its connections from the occurrence of adjacency between these terms. Therefore, observing the variations of network, it is identified how to textual namesakes are related and through an context analyzed how if be there, check whether it is used to express more than one meaning. Results: A generic process with stages of preprocessing, building of a Complex Network used to Natural Language Processing for the building of a network homonyms to extract semantic information textual. Conclusions: The analyze of homonyms selected and labeled is the process not only morphosyntatic, adding semantic in the phrase, paragraph or text where the words are applied. However, with Natural Language Processing an events and philosophical facts can be better analyzed through of a world written textually, for example, the power of argument and the writing of an author profile.