A New Approach to Sentiment Analysis on Twitter Data with LSTM

Sentiment analysis, also known as opinion mining, aims to categorize subjective information in texts into positive, negative, or neutral sentiments. It involves analyzing words to determine emotions expressed in sentences, paragraphs, or documents. The process includes cleaning the text by removing unnecessary characters, stop words, and punctuation, as well as tokenization to divide the text into separate words while maintaining contextual meaning. Analyzing sentiment in Twitter data requires extensive preprocessing, including the removal of unnecessary characters, handling hashtags, and tokenization. Long Short-Term Memory (LSTM), a type of recurrent neural network, is an effective technique for sentiment analysis as it captures sequential and contextual information from text data. In this study, an LSTM model is proposed to perform sentimental analysis on Twitter data. The proposed model was able to achieve 93% accuracy with 0.93 F1 score, outperforming other models.