Regresar

Sentiment Analysis Based on User Opinions on Twitter Using Machine Learning

Abstract:

The growing influence of users on social networks has caused positive and negative content to spread across the Internet. This paper arises from the need to have new tools that allow an analysis of sentiments based on the opinions made by users on the social network Twitter. In particular, the results of identifying the positive and negative opinions issued to a user are presented in order to develop effective communication strategies through social networks. The results obtained allow decisions to be made to obtain a competitive advantage, evaluating positive opinions or, failing that, identifying negative comments to establish strategies to overcome user dissatisfaction. Sentiment analysis can be performed in several ways; however, to obtain greater precision, naïve bayes classifier was used as a machine learning technique, obtaining an accuracy value of 86.2% in the evaluation of the model.