Al-Hams and Al-Jahr Sifaat evaluation using classification approach

Recitation of Al-Quran is an essential task for every Muslim across any background, where Muslims are required to excellently recite Al-Quran that was revealed in Arabic language. To fulfill this task, the first step that is commonly conducted yet takes so much learning time, is to master the pronunciation of the basic unit of the verses which are the letters. In Tajweed knowledge, there is a standard way to pronounce the letters correctly. Makhraj (Points of Articulation) and Sifaat (Classifications) are the first things to be studied for Quranic letters correct pronunciation, where these two topics used to be studied traditionally with series of face-to-face meetings between students and teachers. Up to this date, there are limited research in the literature to examine the correctness of the Sifaat, as an important foundation in Quranic recitation, but research on Quranic verses evaluation is widely available. Therefore, an automated learning system for evaluating Makhraj and Sifaat would be a complementary tool for the students to reduce the time required for learning. As a part of the automated system’s developed, therefore, in this paper, a classification approach is introduced to develop a classification model that can classify the Quranic letters to its first pair of Sifaat with opposites (Al-Hams and Al-Jahr). Mel-frequency Cepstral Coefficients (MFCC) technique was used as features extraction, obtained from the pre-processed audio signal. Features selection technique was then implemented to reduce the size of the features vector, where later, K-nearest Neighbor (KNN) algorithm was used as the classification technique. As a result, it has shown that the reduced features vector has outperformed the full features vector in all KNN methods by 91.1% of accuracy