Key points are not available for this paper at this time.
Sentiment Analysis deals with identifying and aggregating the sentiment or opinions expressed by the users. Sentiment Analysis could be carried out by techniques that may or may not use a lexicon for polarity identification. Training data set i.e. already tagged opinions may or may not be used. Some sentiment analyser could be language dependent or some could be language independent, also called as multilingual sentiment analysers. A detailed survey on different techniques used in Sentiment Analysis is carried out. These techniques are then compared on the basis of usage of lexicon, requirement of training set and language dependency. These are then summarised and analysed. General Challenges are also discussed.
Ghag et al. (Tue,) studied this question.