Sentiment Analysis of Social Media Networks Using Machine Learning

With emergence development of the Web 2.0, there is a huge amount of textual content over the internet including news articles and historical documents, with a notable increase after the rise of social media, such as Twitter platform. More people start to express their feelings and opinions across the internet and various social media. This led to an increase in the number of user-generated sentences containing sentiment information. Investigating new methods to gain different insight into how people feel and respond to different situations is inevitable. This paper compares the performance of different machine learning and deep learning algorithms, in addition to introducing a new hybrid system that uses text mining and neural networks for sentiment classification. The dataset used in this work contains more than 1 million tweets collected in five domains. The system was trained using 75% of the dataset and was tested using the remaining 25%. The results show a maximum accuracy rate of 83.7%, which shows the efficiency of the hybrid learning approach used by the system over the standard supervised approaches.