A pattern-based approach to detect irony in twitter sentiment analysis

Abstract

Twitter sentiment analysis poses challenges due to the informal language, limited character count, and prevalence of sarcasm, which can alter the polarity of messages. This paper presents a pattern-based approach to detect irony in Twitter sentiment analysis. By analyzing various types of irony and identifying their patterns, this paper proposes a methodology to improve the efficiency of sentiment analysis. Tweets are classified into different categories based on their sarcasm using a machine learning algorithm. The proposed approach involves feature extraction from tweets, including sentiment-related features, punctuation-related features, grammatical and phonetic features, and patternbased features. A hybrid pattern extraction with a classification model is employed to process tweet data and classify it as sarcastic or not. Experimental results demonstrate the effectiveness of the proposed approach in detecting sarcasm in tweets, with precision ranging from 84.6% to 98.1% across different classifier algorithms. This pattern-based approach offers promising results for enhancing sentiment analysis on Twitter and understanding the nuances of communication in social media discourse.