The Hyper-Lexi phantom feature selection based on Topic Vector Lex-Sense Convolution Neural Network for Detection of Fake News in Social Media

  • J. Lysa Eben, Dr R. Renuga Devi,
Keywords: Deep Neural network, social media, tweet analysis, lexical term prediction, Convolutional neural network, topic vector, entity content analysis.

Abstract

Current social media contain a lot of information discussion and spreading based on trending hashtags. During the phenomena, the trustworthiness of information leads rumors to produce wrong information spreading over communal social resources. Identifying fake information in tweets/tags is difficult because wrong tags prediction failed in feature selection and classification. Due to unambiguous word frequency in feature selection, the classification failed to predict misleads count terms to reduce the classification accuracy. To resolve this problem, we propose a Hyper-Lexi phantom feature selection (HLPFS) based on Topic vector Lex sense (TVLS) be optimized with a convolution neural network to predict the fake tweets information in social media. Initially, the preprocessing was carried to fake redundant data to predict the topic-relevant model using the Topic vector summarization rate (TVSR). This selects the geometric features of discussion forums content. Based on that topic relevance process, the Reliable entity content hit rate (RECHR) observe positive recurrent information difference between the tweets lexical weights. Hyper-Lexi phantom feature selection (HLPFS) selects the importance of fake difference metrics among reliable terms to reduce dimension based on the entity rate. Then selected features are classified with Adaptive decision in convolution neural network optimized with Relu activation function. The proposed system produces high performance in predicting fake information in social media as well as other methods. The results improve the precision rate as well recall providence to improve the classification accuracy.

Published
2021-09-04
How to Cite
J. Lysa Eben, Dr R. Renuga Devi,. (2021). The Hyper-Lexi phantom feature selection based on Topic Vector Lex-Sense Convolution Neural Network for Detection of Fake News in Social Media. Design Engineering, 5870-5891. Retrieved from http://www.thedesignengineering.com/index.php/DE/article/view/4020
Section
Articles