Análise de sentimento utilizando ferramentas open source

Abstract

Companies are frequently worried about customer’s opinion on their products and offered services, on public interests or on what these people are publishing on social networks. Using the Internet it is possible to get many unstructured data carrying latent information that may give insight on public opinion about a specific subject. These insights, in turn, could be used as base information for the strategic decision-making process. However, the highly available data over Internet is usually unstructured textual information, which is inadequate for the process of pattern extraction and knowledge discovery. In this context, the aim of this paper is to present and apply several text pre-processing techniques on data gathered from social networks and convert them into an adequate input data set for machine learning algorithms.