K-means variations analysis for translation of English Tafseer Al-Quran text

Text mining is a powerful modern technique used to obtain interesting information from huge datasets. clustering distinguish between documents that have the same themes or topics. The absence of datasets ground truth enforces use (unsupervised learning) rather than others, such as classification (supervised learning). “no free lunch” (NFL) theorem supposed no algorithm outperformed other in variety conditions (several datasets). This study aims analyze k-means cluster variations (three algorithms (k-means, mini-batch k-means, and k-medoids) at process stage. Six were used/analyzed chapter Al-Baqarah English translation (text) 286 verses preprocessing Moreover, feature selection term frequency–inverse document frequency (TF-IDF) get weighting term. At final stage, five internal validations metrics implemented silhouette coefficient (SC), Calinski-Harabasz index (CHI), C-index (CI), Dunn’s indices (DI) Davies Bouldin (DBI) regarding execution time (ET). experiments proved k-medoids two terms ET only. In contrast, superior for six datasets, which confirms NFL assumption.