Published Paper Details:

World Wide Web corpus such as social networks, forums, review sites and blogs generate large volume of data in the form of emotions, opinions and arguments about different social events, government policies, political events etc. The social network sites and micro blogging sites are considered a very good source of information nowadays because people share and discuss their opinions about a certain topic freely. With the increased use of technology and social media, people proactively express their opinion through social media sites like Twitter, Facebook, and Instagram. Emotion analysis of social media can help decision makers to understand how people react to policies and events. The unstructured data from the social media is needed to be well-structured for this purpose. This is a challenging task due to the complexity of natural language processing of social media language. Often these messages reflect the emotion, opinion and sentiment of the public through a mix of text, image, emoticons etc. We cannot use standard natural language processing tools to analyze the emotion. People use popular emoticons like smiling face (?), angry face (?) etc., to express emotion instead of words. These statements are often called electronic Word of Mouth (eWOM) and are much prevalent in business and service industry to enable customers to share their point of view. Labeling the emotion as angry, happy or sad is a classification problem. We propose to use a two-step approach to identify the emotion of social media content. Without getting into the complexity of understanding the eWOM, we use machine learning algorithm to solve this classification problem.