The Comparison of Nearest Neighbor Algorithm as Modeling in Conclusion of Interpretation of Bil Ma'tsur and Bil Ra'yi

Tafsir is one of many media utilized to convey the message of the Quran to readers. There are two types of tafsir, namely bil ma'tsur and bil ra'yi, which are grouped based on its respective sources known as shahih and dhaif. The main issue in practicality is differentiating between the two groups is not always easy. The technique most often used in machine learning for classification is the K-Nearest Neighbor (K-NN). On the other hand, K-NN has 3 variations of the algorithm, namely Fuzzy K-Nearest Neighbor (FK-NN), Modified K-Nearest Neighbor (MK-NN), and Improvement K-Nearest Neighbor (IK-NN). This study compares the four algorithms by modeling the best algorithm based on algorithm performance and algorithm complexity. From this research, it was found that MK-NN is a reliable algorithm with an accuracy of 98.07% and a processing time of 2.97 minutes in the deduction of bil ra’yi and bil ma’tsur with 486 documents in Surah Al-Baqoroh and Surah Ali Imron as samples. Furthermore, the best modeling of MK-NN is implemented into a programming language for the application of determining the type of interpretation, with the results of the User Acceptance Test (UAT) and user ease of approaching 99%.