Dual-stream Self-Attentive Random Forest for False Information Detection

Abstract

The prevalence of online social media facilitates massive knowledge acquisition and sharing throughout the Web. Meanwhile, it inevitably poses the risk of generating and disseminating false information by both benign and malicious users. Despite there has been considerable research on false information detection from both the opinion-based and factbased perspectives, they mostly focus on tailored solutions for a particular domain and carry out limited work on leveraging multi-faceted clues such as textual cues, behavioral trails, and relational connection. We propose a novel dual-stream attentive random forest that is capable of selecting clues of discriminative information from individuals, collective information (e.g., texts), and correlations of entities (e.g., social interactions) adaptively. In particular, we use an interpretive attention model for learning textual contents. The model treats the important and unimportant content differently when constructing the textual representation and employs a multilayer perceptron to capture the hidden complex relationships among features of side information. We further propose a unified framework for leveraging the above clues, where we use attentive forests to provide probabilistic distribution as predictions over the two learned representations, which are then leveraged to make a better estimation. We conduct extensive experiments on three real-world benchmark datasets for fake news and fake review detection. The results show our approach outperforms multiple baselines in the accuracy of detecting false information.

Hoax Detection in Indonesia Language Using Long Short-Term Memory Model

“…Various studies on detecting hoax news in Indonesia have been carried out with the use of word similarity measurement theories such as Levenshtein Distance [4]. In addition, research on hoax detection has also been applied using various classical machine learning methods such as K-Nearest Neighbor [5], Decision Tree [6], Naïve Bayes [7], Support Vector Machine [7], Random Forest [8], and so forth [9] [10]. However, the application of classical machine learning methods has weaknesses, including not suitable for processing large and complex data, requires an expert to label data and feature extraction manually, cannot learn based on raw data, and difficulty in representing data.…”

Detecção de notícias falsas usando poucos dados positivos rotulados

“…Algoritmos como K-Nearest Neighbors (k-NN) (COVER; HART, 1967), Naïve Bayes (NB) (LANGLEY et al, 1992), e Extreme Gradient Boosting (XGB) (CHEN; GUESTRIN, 2016), também foram avaliados na detecção de notícias falsas COVÕES, 2020;KALIYAR et al, 2020;REDDY et al, 2020;KHANDELWAL;KUMAR, 2020;SINGH;GHOSH;SONAGARA, 2020;YAVARY;SAJEDI;ABADEH, 2020;ASGHAR et al, 2019;BHUTANI et al, 2019;DONG et al, 2019;MEHTA, 2019 Abordagens que propuseram novos algoritmos também se destacaram na literatura. Em SANTOS; PARDO (2020) foi criado um modelo que reproduz os passos de um usuário realizando uma busca no Google por uma notícia.…”

“…Modelos de aprendizado baseados em redes neurais também foram amplamente propostos, usando técnicas como Multi-Layer Perceptron (MLP), Deep Neural Networks, Recurrent Neural Networks (RNN), Convolutional Neural Network (CNN), Gated Recurrent Unit (GRU), Long Short Term Memory (LSTM) e Transformers para detecção de notícias falsas (CAPUANO et al, 2023;KALIYAR et al, 2020;KANG;HWANG;QAZI;ALI, 2020;WANG et al, 2020;WU et al, 2020a;YAVARY;SAJEDI;ABADEH, 2020;BONDIELLI;MARCELLONI, 2019;MEEL;VISHWAKARMA, 2019;SHARMA et al, 2019;DONG et al, 2019;MONTI et al, 2019;PHILIP, 2019;LI et al, 2019).…”

“…Algoritmos híbridos (ensembles), também foram propostos para detectar conteúdo enganoso. DONG et al (2019) propôs um ensemble baseado em attentive RF e redes neurais profundas. Attentive RF seleciona pistas discriminativas de indivíduos, textos e correlação de interações sociais de forma adaptativa, enquanto RNNs capturam relações ocultas complexas em informações textuais.…”