Social networks have evolved the way we communicate; they have become a new source of information and expression. It is a spontaneous, free, immediate, diverse and universal communication, which allows people to actively participate in topics of their interest. The content generated in the interaction with social networks is vast and has become a valuable source of information, which needs to be analyzed and explored by applying techniques for the classification or labeling of sentiments, in order to find patterns or trends in people’s behavior, which support organizations in strengthening their tasks related to digital marketing. This article takes a theoretical tour of various techniques used to classify feelings on social media. The SVM (Support Vector Machines) supervised learning linear classifier is proposed for the classification or labeling of sentiments in social networks.