Analysis of Drug User Pattern Grouping Based on Social and Psychological Factors Using the K-Modes Method

Abstract

Drug abuse in Indonesia remains a complex social problem with multidimensional impacts on the health, economy, and social stability of the community. South Sulawesi, particularly Palopo City, is an area with increasing rates of drug abuse from year to year. Prevention efforts are often hindered by a lack of information regarding user behaviour patterns, which vary based on social and psychological factors. This study aims to group drug user patterns using the K-Modes algorithm and determine the optimal number of clusters using the Silhouette Coefficient method. The research data is secondary categorical data from the BNNK Palopo City in 2024, with variables including age, education, occupation, social environment, and duration of drug use. The analysis was conducted using R software with the klaR and cluster packages. The results showed that the highest Silhouette value was obtained in five clusters with an average silhouette score of 0.3363, which represents the five main types of drug users based on social and psychological characteristics. Cluster 4 has the highest drug user pattern based on the k-modes analysis. Cluster 4 represents the group of Young Adults Exposed to Strong Social Environments with low educational backgrounds and precarious employment. This permissive social environment makes them susceptible to drug exposure. Interventions that can be provided to users focus on alternative education and the development of life skills.