A NOVEL SECURITY FRAMEWORK FOR PREVENTING ATTACKS IN SENTIMENT ANALYSIS

ABSTRACT

Nowadays, Twitter data-based sentiment analysis is the mainly common topic in Natural Language Processing (NLP). Nevertheless, security attacks on Twitter data are increased day by day because hackers and the attacks will reduce the performance of sentiment analysis. Many kinds of research are developed to overcome this problem, but there are no accurate results found. So this current research proposed a novel Ant Lion honeypot with Regression (ALHR) for detecting the attacks and continuous monitoring of data. Moreover, the fitness function of the introduced replica is used for preventing attacks and continuous monitoring. Also, this model utilizes Twitter-based data about the corona disease 2019 (COVID-19) for detecting attacks and enhances the classification of sentiments using continuous monitoring. For verifying the effectiveness of ALHR technique, launch attacks in classification layer. The developed technique is executed in Python, and the achieved performance metrics are compared with another existing replica regarding the accuracy, recall, precision, F-measure, and error rate. Finally, the ALHR technique enhances the sentiment analysis and provides continuous monitoring. © 2022, Engg Journals Publications. All rights reserved.