| dc.description.abstract | The stories of the prophets and messengers are a fundamental part of Islamic teachings, containing moral, spiritual, and exemplary values for humanity. However, their lengthy and non-interactive presentation often becomes a barrier for people, especially younger generations, to study them in depth. This study aims to develop an interactive chatbot-based Question Answering (QA) system using the Bidirectional Long Short-Term Memory (BiLSTM) algorithm with an Attention Mechanism to enhance understanding of the stories of the prophets and messengers. The system was developed using a dataset of 4.588 question-answer pairs sourced from trusted Islamic literature. The data was divided into training and validation sets. The model was trained for 50 epochs, with the best performance achieved using a batch size of 4, resulting in a validation accuracy of 95% and an mAP50(B) score of 0.98. System testing was conducted on 110 user-submitted questions, with the chatbot correctly answering 104 of them, achieving an accuracy of 95% and an average response time of 132.74 milliseconds. Performance evaluation metrics show outstanding results, with a precision of 100%, recall of 94.5%, and an F1-score of 97.2%. These results demonstrate that the BiLSTM with Attention Mechanism is effective in building an accurate and contextually relevant Islamic QA chatbot. This research contributes to the development of Islamic chatbot technology as a fast, interactive, and reliable learning tool. | en_US |
