Abstract
In recent years, text mining and sentiment analysis have received great attention due to the abundance of opinion data that exist in social networks such as Facebook, Twitter, etc. Sentiments are projected on these media using texts for expressing feelings such as friendship, social support, anger, happiness, etc. Existing sentiment analysis studies tend to identify user behaviors and state of minds but remain insufficient due to complexities in conveyed texts. In this research paper, we focus on the usage of text mining for sentiment classification. Illustration is performed on Tunisian users' statuses on Facebook posts during the "Arabic Spring" era. Our aim is to extract useful information, about users' sentiments and behaviors during this sensitive and significant period. For that purpose, we propose a method based on Support Vector Machine (SVM) and Naïve Bayes. We also construct a sentiment lexicon, based on the emoticons, interjections and acronyms', from extracted statuses updates. Moreover, we perform some comparative experiments between two machine learning algorithms SVM and Naïve Bayes through a training model for sentiment classification.