Comparative Analysis of Deep Learning Models for Sentiment Analysis Using Twitter Data

The overwhelming growth of the WWW has given birth to a new breed of expressing personal opinions. It is an information-intensive platform where one can retrieve other persons' views, organized in various emotional classes, that are becoming increasingly instrumental in decision-making. Organizations, governments, and researchers can explore opinions and perspectives to derive business intelligence and improve decisions. A thorough and detailed comparison of sentiment analysis on Twitter data using CNN, LSTM, and CNN-LSTM is provided in this paper. The applicability of models is evaluated on performance parameters like Accuracy, Precision, Recall, and F1-Score. This research gives vital insights into the benefits and drawbacks of these algorithms by analyzing their efficacy on a larger dataset like Twitter. It also helps professionals decide which of the models fits best into their needs.