Sarcasmometer using sentiment analysis and topic modeling

Sarcasm is a type of sentiment where the notion that is conveyed is contrasting to what it actually means. It is usually used to denote something that is funny or to denote anger or dislike regarding a particular situation. Sarcasm has become a part of our daily lives. It is used in various social networking sites, review posts, entertainment businesses, etc. However, different people have different interpretation of sarcastic texts and this leads to debatable opinions about the product which is being described. Recognizing sarcastic statements can be very useful as it enhances the efficiency of after-sales services or consumer assistance through understanding the intentions and real opinions of consumers when browsing their feedbacks or complaints. In this paper we propose a system that will measure sarcasm using tweets from Twitter. We propose different algorithms to calculate the effect of sarcasm on texts and generate a score. Different features are generated from the received tweets which helps us to generate the score. At the end, we compare the scores from different algorithms to present the most efficient way to detect sarcasm. The system also provide a separate portal to check the score of any sentence/text entered by user and determine its score using the most accurate algorithm.