Automated evaluation of hadith narrator reliability using machine learning and rule-based models based on Al-Jurh and Al-Ta’dil literature

The integration of machine learning and rule-based methodologies grounded in Al-Jurh and Al-Ta'dil literature offers an effective approach to improving the accuracy and objectivity of Hadith narrator reliability assessment by enabling the analysis of large-scale biographical and Hadith data to identify complex patterns beyond conventional human analysis. This study explores how machine learning techniques can be applied to assist in the authentication of Hadith reporters, which is an important step in the process of documenting Hadith in the study of Islam. The discipline of Al-Jurh and Al-Ta’dil in Hadith criticism provides a systematic framework for evaluating the reliability of the narrators based on historical and biographical arguments. The current research paper investigates the use of machine learning algorithm to analyze available data on existing Hadith criticism data and identify generalized patterns and predict regarding the reliability of the narrator, which can potentially significantly streamline the process of documenting and producing new findings. Moreover, the study will evaluate the cost-effectiveness and accuracy of using machine learning to automate or supplement conventional methods of Hadith criticism, potentially helping to bring about the development of a more effective, data-driven paradigm in the context of Hadith studies. The ability of the machine learning methods to help in the verification of Hadith reporters and to derive complex patterns and interrelations within the texts is supported by experimental results. Notably, the rule-based approach achieved the highest performance, attaining 98% classification accuracy and significantly outperforming other machine learning models. The accuracy of the Random Forest algorithm was 84%, and that of a naive Bayesian classifier was 78%. These results further demonstrate the effectiveness of the rule-based methods in situations when expert knowledge is formally encoded and applied to make the accurate predictions. However, the performance of data-driven models, such as the Random Forest and the Naive Bayes classifier, indicates their practical potential, particularly in scenarios involving larger and more linguistically diverse datasets.