dc.description.abstract | Topics searching in the Holy Quran using word derivative is difficult task without arabic word morphology knowledge. Light stemming is stem extraction method that can replaces word derivative plays in topics searching. However, this method fail to removes infix so clustering method can be better hopefully. This research aims to build word derivative searching system using light stemming and clustering that appropriate with Indonesian speaker’s query. There are two types of query that accepted by the system that have 30 queries in each: Latin and Arabic script. Searching quality evaluation using Average Precision (AVP), Mean Average Precision (MAP), and Precision at N shows that light stemming is better than clustering with 0.30 AVP in light stemming and 0.03 AVP in clustering. Whereas for running time, light stemming is four time faster than clustering at Arabic script. | en |