An efficient FTS-BERT based fake news detection using CKH_GANs classification technique
Pooja Malhotra,Sanjay Kumar Malik
DOI: https://doi.org/10.1007/s11042-024-19249-x
IF: 2.577
2024-05-03
Multimedia Tools and Applications
Abstract:The way by which the news is created and published has been changed by the appearance and the extensive adoption of the Semantic Web (SW) and the concepts of Social Media together with the growth of the Internet. The spread of Fake News (FN) is also prevalent. However, the recognition of FN from the disparate sentences of various website data becomes a challenging task. Hence, an FN detection system is developed in this paper by employing the proposed CKH_GAN model. The process is initiated by gathering data from the scraped dataset, which contains different news and tweets. Then, the acquired data is pre-processed to remove the repeated details, hashtags, and meaningless usernames. Further, the redundant website data are eliminated using the HDFS (Hadoop Distributed File System)-Map Reduce approach. Subsequently, the disparate characteristics, including grammar less and factual less sentences are detected and grouped based on the similarity measures by using the developed FTS-BERT (Fined Tuned Scaling-centered Bidirectional Encoder Representations from the Transformer) algorithm. Further, the informative features are extracted, and the significant features among them are chosen utilizing the proposed CP-EO (Cutting Plane method-centered Equilibrium Optimizer) technique. Eventually, based on the selected features, the FN and real news are categorized using the proposed CKH_GAN (Calculus Krill Herd based Generative Adversarial Networking (CKH_GANs) model for identifying the FN. The experimental outcomes exhibit that the proposed work detected FN with an accuracy of 95.89%, precision of 93.65%, f-score of 94.87%, FPR of 11.23%, and FNR of 10.98%, which are better than the other existing techniques.
computer science, information systems, theory & methods,engineering, electrical & electronic, software engineering