The Combination of Agglomerative Clustering and K-Means Algorithm for Website Visitor Segmentation
The Combination of Agglomerative Clustering and K-Means Algorithm for Website Visitor Segmentation
Author : YUDHA AGUNG WIRAWAN; INDWIARTI; YULIANT SIBARONI Published on : GTAR Conference 2015
Abstract
Clustering is an important aspect in the use of website mining for web visitor segmentation. In this writing, we classify visitors using hierarchical and non-hierarchical clustering method on academic website log data. Hierarchical and non-hierarchical classification methods employed in this final project are Agglomerative Clustering and K-Means. While Agglomerative Clustering is used to determine the total number of cluster, K-Means is employed for the segmentation. According to the grouping results and detailed observation on webpage accessed by each cluster, it is found that from 2 clusters, menus accessed by users are registration menus and student academic menus. Furthermore, the most frequently accessed registration menus include fee payment invoice, late registration and registration process. On the other hand, most often accessed student academic menus are schedules, attendance, and syllabus. Menu services in those two clusters are accessed in the period of student registration and at the beginning of study term. From 3 clusters, menu service accessed by users is academic menu service. The most frequently accessed menus in this cluster are registration process and final project. Menu services in those 3 clusters are accessed in the period of study plan amendment.
