2026· International journal of research and scientific innovation· 0 citations
TL;DR
HEF-XFND is proposed, a hybrid explainable feature-fusion framework that combines sparse lexical evidence, contextual transformer representations, source-level credibility indicators, and calibrated ensemble learning that addresses three recurring limitations in fake-news research.
Abstract
The unchecked diffusion of fabricated and misleading news across digital platforms has created a need for detection systems that are accurate, interpretable, and deployable at scale. This paper proposes HEF-XFND, a hybrid explainable feature-fusion framework that combines sparse lexical evidence, contextual transformer representations, source-level credibility indicators, and calibrated ensemble learning. The architecture extracts term frequency-inverse document frequency features, linguistic style descriptors, and BERT/RoBERTa sentence embeddings. These heterogeneous representations are projected into a common feature space and processed by complementary classifiers, including logistic regression, linear support vector machines, random forests, XGBoost, and a lightweight neural classifier. A stacking layer produces the final veracity probability, while SHAP- and LIME-based explanations identify the words and feature groups that most strongly influence a decision. The framework also includes probability calibration, confidence scoring, crossvalidation, and latency-aware model selection for high-volume deployment. A reproducible experimental protocol is presented for the Fake and Real News dataset containing 44,898 articles. Literature-informed illustrative results indicate that transformer-assisted ensemble fusion can outperform isolated conventional and neural baselines while retaining practical inference efficiency. The proposed design addresses three recurring limitations in fake-news research: dependence on a single representation, insufficient explanation of predictions, and inadequate consideration of operational scalability. The manuscript provides a transparent foundation for subsequent implementation, external validation, and deployment in newsrooms, fact-checking services, and social-media moderation systems.
A hybrid transformer-based ensemble model for automated fake news identification using the FakeNewsNet dataset is proposed and Experimental results show that the ensemble model achieves an accuracy of approximately 93%, outperforming the individual constituent models.
M. E. Babu, G. Sukanya· International Journal for Re...· 0 citations
Publication of fake/false news on social media exceeds the controlled ability of manual fact-checking, and the labeling of operations is often not complete and trustworthy. In this research, We explore twitter level truthfulness classification on Truthseeker 2023 ground truth corpus, and use a hybrid feature representation that incorporates textual features as TF-IDF(Term Frequency-Inverse Document Frequency) vector, together with pre-calculated user, content and engagement features. The baseline experiments include initial experiments of standard supervised models, such as, Logistic Regression, Random Forest, Support Vector Machine, XGBoost, and LightGBM as well as a soft-voting ensemble, on a subset of a held-out test split. These findings show that linear and boosted -tree models have strong discriminatory capability, and the ensemble achieved an accuracy of 93.73 and ROC-AUC of 95.39. We then learn using the label scarce regime through Positive Unlabeled (PU) learning where a fraction of the target class is labeled and the rest are explicitly unlabeled mixture. In the PU model, Elkan-Noto approach presents the highest level of efficiency, with an accuracy of 84.53 per cent and an ROC-AUC of 90.40; a Two-step PU strategy presents a higher- recall option. All in all, the results measure the performance loss to be expected in passing between fully supervised models to PU learning and show that even with the necessary calibration of PU strategies (semi-supervised learning), a significant percentage of the intrinsic signal can be recovered without the need to explicitly label examples as negative.
The speed with which fake news is being spread through digital channels has presented a great challenge, impacting the public opinion, politics, and social cohesion. Several machine learning and deep learning methods have been suggested for detecting fake news, but most of these models are either highly resource-intensive or do not give insights into the news predictions. In response to these challenges, this paper presents an Explainable Artificial Intelligence (XAI) framework for fake news detection that unites the complementary explainability methods: SHapley Additive exPlanations (SHAP) and Local Interpretable Model-agnostic Explanations (LIME) with the feature extraction technique, Term Frequency-Inverse Document Frequency (TF-IDF) and the Linear Support Vector Machine (Linear SVM) classifier. The framework is based mainly on the GossipCop and PolitiFact subset of the FakeNewsNet repository. The methodology is divided into four phases, namely: data preprocessing, textual feature extraction using TF-IDF, classification using Linear SVM, and explainability analysis of the classification by using both local and global interpretation techniques. The experimental results confirmed the proposed approach with accuracy of 79.40%, precision of 80.51%, recall of 95.34%, F1-score of 87.30%, and ROC-AUC of 81.40%, and by providing transparent explanations of classification using SHAP and LIME. The achieved outcome shows the proposed framework has a good balance between classification accuracy, simplicity of calculation, and the ease of understanding the model. As a result, the proposed approach is an effective solution for detecting fake news in applications, including automated fact-checking, digital journalism and social media content monitoring.
The rapid proliferation of digitally distributed news has made large-scale automated verification an urgent research priority, as deceptive content spreads across social platforms faster than manual reviewers can evaluate it. This work introduces a two-stage deep learning pipeline in which a pretrained BERT encoder generates rich contextual token embeddings that are subsequently processed by a Bidirectional Long Short-Term Memory (BiLSTM) network, enabling the classifier to exploit both sentence-level semantics and document-level narrative flow simultaneously. Experiments on the WELFake benchmark [1] following a standardised cleaning pipeline—lowercasing, URL stripping, tokenisation, stop-word removal, and WordNet lemmatisation— yielded 98.7 % accuracy and an F1-score of 0.986. This surpasses a BERT-only baseline by 3.1 percentage points and exceeds several previously published state-of-the-art results. The gains confirm that coupling transformer-based semantic representations with recurrent sequential modelling produces a measurable and reproducible improvement in misinformation detection. Planned extensions include multilingual evaluation and knowledge-distilled encoder replacements targeting real-time throughput constraints.
Ardra P Namboodiri, Archa P S, Honey Mol O· 2026 6th International Confe...· 0 citations
ABSTRACT
The rapid proliferation of fake news and misinformation across digital platforms poses a serious threat to democratic processes, public health, and social stability. This research paper presents a comprehensive study of an AI-based fake news detection system leveraging Natural Language Processing (NLP) techniques and multiple machine learning algorithms. The proposed system analyzes textual content, linguistic patterns, writing style, and sentiment cues to automatically classify news articles as real or fake.
Several classification algorithms — including Naive Bayes, Logistic Regression, Support Vector Machine (SVM), Random Forest, Decision Tree, and Long Short-Term Memory (LSTM) neural networks — are implemented and compared. Feature extraction is performed using TF-IDF and Bag of Words representations. The system is evaluated using standard metrics: Accuracy, Precision, Recall, F1-Score, and Confusion Matrix.
Experimental results demonstrate that deep learning models, particularly LSTM, achieve superior contextual understanding, while traditional models offer faster inference. The paper also discusses current challenges such as sarcasm detection, multilingual content, and computational constraints, along with future directions including real-time monitoring and transformer-based architectures. This work contributes significantly to the broader effort of combating misinformation through technological innovation and provides a comprehensive roadmap for future research in this critical area.
Keywords: Fake News Detection, Natural Language Processing, Machine Learning, Deep Learning, LSTM, TF-IDF, Text Classification, Misinformation, Social Media
Shahid Khan, Dr. Abdul Majid Farooqi· International Scientific Jou...· 0 citations
The rapid proliferation of user-generated content on social media platforms has transformed the way information is created, shared, and consumed. While this democratization of information has clear benefits, it has also enabled the large-scale dissemination of fake news, which poses serious threats to public health, political stability, financial markets, and social cohesion. Automatic fake news detection has therefore become an active and urgent research problem in computer science. This paper proposes a multi-model learning framework that combines the complementary strengths of classical machine learning classifiers, deep sequential neural networks, and transformer-based contextual language models to detect fake news on social media. The framework integrates lexical, semantic, and contextual features through a weighted soft-voting ensemble that fuses predictions from a Support Vector Machine, a Bidirectional Long Short-Term Memory network, and a fine-tuned BERT encoder. Experiments were conducted on three widely used public benchmark datasets, namely LIAR, FakeNewsNet, and ISOT, comprising more than one hundred thousand labeled news statements and articles. The proposed ensemble achieved an accuracy of 96.4% and an F1-score of 96.1% on the ISOT dataset, outperforming each individual constituent model and several competitive baselines reported in the literature. Ablation studies confirm that the transformer component contributes the largest performance gain, while the ensemble fusion improves robustness and reduces variance across datasets. The results demonstrate that combining heterogeneous learning paradigms yields a more accurate and generalizable fake news detector than any single model in isolation. The paper also discusses computational cost, interpretability, and limitations, and outlines directions for multimodal and cross-lingual extensions.
Mrs Priyal Verma· International Journal of Res...· 0 citations