In recent years majority of research is carried in the arena of opinion mining particularly the textual data which is available on the social media. Sentiment analysis is extensively used in online reviews, social media and various applications which extend from advertisement to consumer service. It is used to obtain a clear view of the attitudes, sentiments and sensations of individuals which is conveyed in social media. It is a process of defining whether the user’s blogs is positive, negative or neutral. Sentiment analysis has many challenges and the most important is the detection of sarcasm. The classification of the type of sarcastic sentences is a perplexing task. In this work, a supervised classification technique i.e. Multinomial Naïve Bayes (MNNB) is used to detect sarcasm and SVM (support vector machine) is used to detect the sarcasm types. In this paper, the sarcasm is extracted from the tweets by means of MNNB. The tweet contains noisy messages and it has been handled well for effective recognition of sarcasm. In addition, the type of sarcasm also identified in order to diagnose the mood of the user.
