<![CDATA[The ability to read the hijaiyah letters correctly is an important foundation in learning the Qur'an. However, the low Qur'an literacy in Indonesia indicates the need for technological innovation to support the learning process. This article presents a systematic literature review on the application of the Convolutional Recurrent Neural Network (CRNN) algorithm and the Mel-Frequency Cepstral Coefficients (MFCC) feature extraction technique in speech classification, with a focus on their potential implementation for recognizing the pronunciation of the hijaiyah letters. The analysis was conducted based on ten relevant studies selected using the PRISMA method. The results of the study indicate that MFCC is effective in representing the phonetic characteristics of sounds, including in the Arabic context. Meanwhile, CRNN has proven superior in managing audio data with tempo and sequential structure. The combination of the two has strong potential to build an accurate and adaptive hijaiyah letter sound classification system, especially in supporting speech-based tajweed learning. This study provides a conceptual basis for the development of an artificial intelligence-based Qur'an learning application.]]>
Copyrights © 2025
