Sentiment Mining from Text: A Technical Review

Expression of sentiments is an integral aspect of human communication. The prevalence of Internet and social media platforms has facilitated the exchange and storage of large volumes of human communication in the form of natural language text which can be mined for various intelligencebased applications. In order to build computers that can serve humans better, there are ongoing efforts in computer science research to develop machine learning algorithms that can process the textual data and perform sentiment mining tasks such as, ‘detecting the presence of emotion in text’, ‘selecting a model for representing emotion’, ‘classifying the sentiment polarity of text’ and ‘measuring the intensity of the expressed sentiment in text’. This paper aims to discuss the challenges posed to the various tasks associated with mining sentiment from text, review the existing methods to address these challenges in contemporary literature and identify important areas for future scope of research in this field.