Measuring sentiments in online social networks

Sentiment analysis has being used in several applications including the analysis of the repercussion of events in online social networks (OSNs), as well as to summarize public perception about products and brands on discussions on those systems. There are multiple methods to measure sentiments, varying from lexical-based approaches to machine learning methods. Despite the wide use and popularity of some those methods, it is unclear which method is better for identifying the polarity (i.e. positive or negative) of a message, as the current literature does not provide a comparison among existing methods. This comparison is crucial to allow us to understand the potential limitations, advantages, and disadvantages of popular methods in the context of OSNs messages. This work aims at filling this gap by presenting a comparison between 8 popular sentiment analysis methods. Our analysis compares these methods in terms of coverage and in terms of correct sentiment identification. We also develop a new method that combines existing approaches in order to provide the best coverage results with competitive accuracy. Finally, we present iFeel, a Web service which provides an open API for accessing and comparing results across different sentiment methods for a given text.