Optimization of the Number of Clusters in Grouping Characteristics of Hajj Pilgrim Candidates in Tebing Tinggi City Using the K-Means Algorithm and Elbow Method

Abstract

Every year, thousands of prospective Hajj pilgrims from Tebing Tinggi City register and participate in manasik (Hajj rituals) guidance. The diversity of their characteristics, including age, gender, education, occupation, and frequency of guidance, demands different service strategies. This study aims to group prospective Hajj pilgrims based on demographic characteristics using the K-Means algorithm with cluster number optimization via the Elbow method. The data used is the registration data of prospective Hajj pilgrims for 2023 from KBIHU in Tebing Tinggi City, totaling **189 records**. The variables used include age, gender, education, and occupation. The results show that the optimal number of clusters is **K=3** with an inertia value of 152.47 at K=3. Cluster 1 (Elderly) is dominated by pilgrims aged over 50 years with secondary education and occupations as housewives/retirees. Cluster 2 (Productive Adults) consists of pilgrims aged 30-50 years with higher education and occupations as civil servants/private employees. Cluster 3 (Youth) comprises pilgrims under 30 years old with student status. This study provides recommendations for different guidance strategies for each cluster: an intensive approach with visual/audio media for Cluster 1, flexible blended learning for Cluster 2, and independent digital-based guidance for Cluster 3.