3. Research Full Paper
Viviani Kwecko1 , Fernando Tolêdo1, Sam Devincenzi2, José Oxlei1, Silvia Botelho1

This research presents a full paper in which we identify and systematize how the vertiginous growth of social media allows the monitoring of public opinions, with a special focus on analyzing the feelings of the population's opinionated arguments about Education. 

Brazil has a System of Social Perception Indicators (Sips) developed by the Institute for Applied Economic Research whose objective is to capture the opinion and evaluation of the population on public policies. Sips-Educação, released in 2010, outlined the population's conceptions using questions applied across the country. 

A decade after this evaluation, little is known about the real meaning of Education for the population, which prevents an effective mobilization of society in the sense of participating in the educational process, inspecting the quality of teaching and valuing the school, teachers and practices innovative.

 Information and Communication Technologies have generated a large number of data, however despite the enormous amount of information produced, we face difficulties in finding relevant content. Identifying, monitoring and allocating a large volume of opinionated text is a complex task due to: the multiplicity of sites on which these "lines" are expressed; the difficulty in identifying the relevant themes of the opinions; the diversity of the information in the multiple comments. In addition to these structural obstacles, it is necessary to consider that all human analysis of information is subject to interpretive biases. Automated systems of classification, opinion mining and text summarization therefore represent potential tools for the process of understanding the opinionated content of social networks, as they allow for the structuring of data and overcoming some of the subjective biases. This whole range of public opinions is being used for decision making, but still in a limited way to monitor the presence of value in public policies. In this sense, we present a model for the extraction and organization of information oriented to opinion. We have brought together different methods in order to produce better results for the classification and summarization of various documents considering education as the basis of analysis. The proposed model is based on the steps of i) classification of patterns based on Deep Learning; ii) analysis of contexts and visualization of different associative paths in publications through the Implicative Statistical Analysis; and iii) validation of opinion abstracts.And the results presented in this study refer to the database made up of 42,062 publications related to the city. Then 1138 publications were noted. This annotation base considered one or more aspects of the Connected Smart Cities Ranking. With regard to accuracy, we reached a total of 76% for the post classifier and 80% for the sentiment classifier in the sample used during training. These are satisfactory values, and justified by the number of samples that were given as input to the network. The collective social discourses, resulting from the analysis of the summarize the opinions of 820 posts that presented representative terms for the education axis in the negative polarity, of the total of 975 posts classified by the dataset.