Özet:

Abstract In the recent era of social networking, the number of users and amount of data on social network increase rapidly day by day. Any event or activity happened in surrounding people post their feeling and comments about the event or activity on social media. Any new product launched then people also give comments on that product using social media platform. Sophisticated methods of expressing different opinions make it difficult to determine the true state of emotions. People use words to express their negative feeling in positive way called as sarcasm. These sarcastic statements are difficult to understand and very complex to identify by machines. Identification of context of text is useful while detecting the sarcasm from text. In this paper we discuss the new approach, Fuzzy logic-based context identification mechanism (FBCIM) to find the context of the tweets and that context is being used for sarcasm detection using different existing sarcasm detection techniques. FBCIM uses four features extracted from the tweets collected from the tweeter API and using the linguistic information of the four features rule-based evaluation is carried out. Experimentation shows that the FBCIM approach guarantees flexibility and also energy efficient. FBCIM approach is scalable too as increasing the number of tweets does not affect the functioning and performance. Result shows that FBCIM identify the context of text accurately when provided with different datasets which contains the balanced and imbalanced data as well. 

Anahtar Kelimeler: