Enhancing Social Media Sentiment Analysis through Blockchain-Enabled Validation
Sandip A. Kahate, Ganesh Shivaji Pise, K B Queen H Pawar, Nilesh D. Navghare · 2024
This research explores the use of blockchain technology to enhance the accuracy and integrity of sentiment analysis on online social networks. Traditional text-based sentiment analysis methods utilizing machine learning and deep learning have faced challenges in addressing the growing prevalence of abusive, biased, or offensive content on social media platforms. To address this issue, the researchers propose integrating an LSTM framework with a blockchain layer to establish a cryptographic link that allows the transfer of ledger metadata to the learning dataset. This "Proof of Learning" blockchain layer serves as a validation mechanism, regulating the linked data and ensuring the integrity of the sentiment analysis process. The proposed strategy was evaluated using tweets with negative, neutral, and positive sentiment, achieving an accuracy of 92.98%, a significant improvement over traditional techniques, which had an accuracy of only 86% on the same data. The findings highlight the potential of blockchain-enabled validation to enhance the reliability and trustworthiness of sentiment analysis on online social networks.