Holy Quran recitation recognition refers to the process of identifying the type of recitation, among those authorized styles of recitation (Riwayah in Arabic). Several previous studies investigated the recitation rules (Ahkam Al-Tajweed in Arabic) that are applied by readers or reciters while reading the Holy Quran aloud, but no study has examined the problem of tracking the type of recitation used in the reading. Through this research, we can assist Holy Quran students to easily learn the perfect and accurate recitation by successfully applying Ahkam Al-Tajweed and helps them to distinguish between different recitations or "Riwayah". In this paper, a recognition model is conducted to recognize the Riwayah from the corresponding Holy Quran acoustic wave. This model was built upon three phases; the first phase is the Mel-Frequency Cepstrum Coefficients (MFCC) feature extraction of the acoustic signal and labeling it, the second phase is training Support Vector Machine (SVM) learning model the labeled features, and finally, recognizing Riwayah based on this trained model. To attain this, we have built our corpus, which has 10 categories, each of which is labeled as one type of Holy Quran recitation or Riwayah. Different machine learning algorithms were applied and compared. Experimental results proved the superiority of our proposed SVM-based recognition model for Riwayah over other machine learning algorithms by a success rate of 96%
Key words: Arabic language, Quran recitation, Artificial Neural Network, Support Vector Machine (SVM),MFCC.