Paper Contents
Abstract
Text based social media platforms such as Twitter have become hotbeds of extreme ideology, with users expressing their opinions. To preserve social peace and internet safety, extremist content must be identified and removed. Conventional approaches to detecting members of extreme groups frequently focus just on textual analysis, ignoring the important emotional context that is ingrained in the language. This study suggests a novel method for identifying extremist Twitter members by combining language and emotions using deep learning algorithms. The methodology entails preprocessing textual data from Twitter posts using Natural Language Processing (NLP) techniques. This stage of preprocessing involves stop word removal, stemming, and tokenization. Sentiment analysis methods like the VADER (Valence Aware Dictionary) lexicon, which assigns polarity ratings to each tweet and captures the underlying emotional tone, are then used to extract the text's emotional content.The fundamental idea of the suggested approach is to use deep learning architectures to integrate textual and emotional characteristics. To capture syntactic and semantic information, Long Short-Term Memory (LSTM) networks are specifically used to learn the sequential patterns found in the textual material. Simultaneously, the textual characteristics and the emotional data retrieved via sentiment analysis are concatenated to provide the model more context. Using supervised learning, the integrated model is trained by using labelled datatweets connected to extremist affiliatesto optimize the model's parameters. Using linguistic and emotional signals, the model is taught to identify patterns suggestive of extremist sympathies. Extensive experiments on real-world Twitter data labelled with extremist associations are used to evaluate the suggested approach. Performance measures including recall, F1 score, and precision are used to estimate how well the model identifies extremist affiliates.
Copyright
Copyright © 2024 PRABHJOT SINGH. This is an open access article distributed under the Creative Commons Attribution License.
