Using a Novel Hybrid Krill Herd and Bat based Recurrent Replica to Estimate the Sentiment Values of Twitter based Political Data

Abstract

Big data is an essential part of the world since it is directly applicable to many functions. Twitter is an essential social network or big data replicating political information. However, big data sentiment analysis in opinion mining is challenging for complex information. In this approach, the Twitter-based political datasets are taken as input. Furthermore, the sentiment analysis of twitter-based political multilingual datasets like Hindi and English is not easy because of the complicated data. Therefore, this paper introduces a novel Hybrid Krill Herd and Bat-based Recurrent Replica (HKHBRR) to evaluate the sentiment values of twitter-based political data. Here, the fitness functions of the krill herd and bat optimization model are initialized in the dense layer to enhance the accuracy, precision, etc., and also reduce the error rate. Initially, Twitter-based political datasets are taken as input, and these collected datasets are also trained to this proposed approach. Moreover, the proposed deep learning technique is implemented in the Python framework. Thus, the outcomes of the developed model are compared with existing techniques and have attained the finest results of 98.68% accuracy and 0.5% error.