Analysis Partition Clustering and Similarity Measure on Al-Quran Verses

Clustering text is an important application in data mining. This is related by grouping similar text documents together. In this study, several models are builts to classify Qur'anic verses on Surah Al-Baqarah using three clustering technique: k-means, bisecting kmeans, and k-medoid. Every verse in Surah al-Baqarah represented as a document derived from the translation of the Qur'an in English. Three similarity measures are also used: cosine similarity, jaccard similarity, and correlation coefficient. Then, the cluster of each combination of clustering technique with similarity measure is evaluated using average within cluster distance and davies bouldin index. The result show that the best performance is achieved by using the hemodoidal combined with cosine similarity. Finally obtained the category verses in the Surah al-Baqarah that correlate with each other.