Intelligent web archiving and ranking of fake news using metadata-driven credibility assessment and machine learning

The explosive growth of online news has intensified concerns over preserving authentic content and combating misinformation. While established preservation frameworks such as the Open Archival Information System (OAIS) and the Web Archiving Life Cycle Model (WALCM) ensure structural durability and access, they lack mechanisms for evaluating the credibility of news before archiving. This critical gap risks embedding misinformation into permanent repositories, thereby distorting historical records and misleading future users. To address this limitation, we introduce the Enhanced Web Preservation Framework (EWPF), which embeds credibility assessment directly into the digital archiving process. EWPF integrates literature-based identification of fourteen key Credibility Factors (CFs), their weighting through the Analytic Hierarchy Process (AHP), and the development of a custom dataset of 800 annotated articles. Multiple machine learning models were trained for classification, with Random Forest and Gradient Boosting achieving the highest accuracy (0.99). An ablation analysis further revealed the relative impact of each credibility factor, offering deeper insight into the framework's robustness. To promote transparency and user engagement, EWPF also includes an interactive interface that visualizes preservation outcomes and dynamic trust scores. Experimental evaluations confirm that EWPF enhances archiving accuracy while remaining computationally efficient and scalable. By linking preservation with credibility analytics, EWPF provides an empirically validated solution for strengthening trustworthy digital journalism in the era of misinformation.