Automating Sanad Continuity Verification in Disconnected Hadith Using Machine Learning

Hadiths are narratives of the deeds and sayings of the Prophet Muhammad (peace be upon him), transmitted orally by a succession of narrators prior to being written down. In Hadith studies, confirming the veracity of these narrators is crucial since the Sanad, or chain of transmission, is what determines if a hadith is weak (Da'if) or authentic (Sahih). Conventional techniques for determining this authenticity need hadith scholars to manually examine the relationships and lives of narrators, a tedious and error-prone process. This work suggests an AI-driven method for easier analyzing Sanad's continuity that makes use of conventional machine learning models. By comparing these methods, we found that traditional AI models like Hist Gradient Boosting Classifier and Random Forest Classifier achieved higher accuracy, reaching up to 93.16% and 96.55%, respectively, and not only outperformed deep learning models but also demonstrated superior accuracy and stability compared to advanced AI-generation models such as ChatGPT, which achieved an overall accuracy of 77.23%. Our findings demonstrate that the AI-based method outperforms both generative models and advanced AI, while also greatly increasing efficiency, lowering human error, and giving scholars a trustworthy tool for Hadith authentication.