Sentiment Analysis of Tweets Containing Fuzzy Sentiment

Abstract: Sentiment analysis involves determining the polarity of statements—classifying them as positive, neutral, negative, satisfied, or frustrated—based on opinions, feedback, and comments shared on social media or e-commerce platforms. It is a key area in text mining for analyzing and classifying user sentiments. This work focuses on identifying the sentiments of tweets and retweets on Reddit by searching for specific keywords and evaluating their polarity as positive or negative. Sentiment analysis is performed using feature selection for scoring words, with classification algorithms employed for training, testing, and sentiment evaluation. Performance metrics such as accuracy, precision, and time are compared across ANN, CNN, and LSTM models.

Keywords: Fuzzy sentiment, tweet sentiment analysis, sentiment polarity