Evaluating the effectiveness of publishers’ features in fake news detection on social media

Abstract

With the expansion of the Internet and attractive social media infrastructures, people prefer to follow the news through these media. Despite the many advantages of these media in the news field, the lack of control and verification mechanism has led to the spread of fake news as one of the most critical threats to democracy, economy, journalism, health, and freedom of expression. So, designing and using efficient automated methods to detect fake news on social media has become a significant challenge. One of the most relevant entities in determining the authenticity of a news statement on social media is its publishers. This paper examines the publishers’ features in detecting fake news on social media, including Credibility , Influence , Sociality , Validity , and Lifetime . In this regard, we propose an algorithm, namely CreditRank, for evaluating publishers’ credibility on social networks. We also suggest a high accurate multi-modal framework, namely FR-Detect, for fake news detection using user-related and content-related features. Furthermore, a sentence-level convolutional neural network is provided to properly combine publishers’ features with latent textual content features. Experimental results show that the publishers’ features can improve the performance of content-based models by up to 16% and 31% in accuracy and F1, respectively. Also, the behavior of publishers in different news domains has been statistically studied and analyzed.

Fake News Detection System Using Featured-Based Optimized MSVM Classification

“…As we can see in fig. (11) that, our proposed model is underperforming for one model, and it has 97.7% accuracy, 97.8% precision, 96% recall, 96.3% F1-Score almost 1.1% lesser accuracy than the highest accuracy claim by the best algorithm which is 98.8% accuracy, 98.5% precision, 96.6% recall, 97.5% F1-Score in this dataset [38]. Our model is more consistent in every epoch we run while testing it on this dataset, but it cannot get the best accuracy compared to the existing model, which we can improve in features.…”

“…Fake news detection is divided into three parts of detection first is textual data, the second is image-related data, and the third is video-related data. There are lots of models used to detect fake news from all of these outcomes like in [19] blockchain and Bi-LSTM is used to achieve the highest accuracy in all of the given research articles for the dataset of PolitiFact, in Gossipcop which is a smaller dataset the CNN approach used by [38] is excellent in all the other research articles. On Twitter election dataset XLM-RoBERTa CNN approach is used in [50] is giving maximum accuracy among all the other models.…”

A comprehensive survey on online social networks security and privacy issues: Threats, machine learning‐based solutions, and open challenges

“…Karimi et al 187 proposed a fake news detection approach (MMFD) using an automated feature extraction method that can accept the news that comes from multiple sources and detect whether news is fake or real and classify them depending on their degrees of fakeness. Hosseinimotlagh and Papalexakis 188 used the tensor factorization, Wang et al 189 195 used some evaluation models such as count vectorizer, N-gram, and TF-IDF vectorizer with the ML algorithms, Kanagavalli et al 196 used classification based on bi-directional long short term memory (BiLSTM), Jarrahi and Safari 197 introduced convolutional neural network in sentence-level, many other authors [198][199][200][201] also used various ML based approach for detecting fake news on OSNs.…”

“…Using deep convolutional layers, Tembhurne et al 193 developed a multi‐channel model (Mc‐DNN) that extracts the most important features without considering hand‐crafted features. Choudhury and Acharjee 194 used SVM, random forest, naive Bayes, and logistic regression classifiers, Raja and Raj 195 used some evaluation models such as count vectorizer, N‐gram, and TF‐IDF vectorizer with the ML algorithms, Kanagavalli et al 196 used classification based on bi‐directional long short term memory (BiLSTM), Jarrahi and Safari 197 introduced convolutional neural network in sentence‐level, many other authors 198‐201 also used various ML based approach for detecting fake news on OSNs.…”

EFND: A Semantic, Visual, and Socially Augmented Deep Framework for Extreme Fake News Detection

“…Few studies [2,[7][8][9][10][11][12][13][40][41][42][43] have tried to employ user profile characteristics for fake news detection. Wu et al [12] identified bogus news by employing an LSTM network along propagation pathways and obtaining user personal information included from social media.…”