DBSCAN and KMeans Clustering: Comparing Strengths and Suitability in Customer Segmentation

K-Means clustering and Density-Based Spatial Clustering of Applications with Noise, both unsupervised clustering approaches have many applications. This paper performs a comparative analysis of the suitability and strengths of these two methods in segmenting customers by emphasizing profession, annual income, spending score, family size, and work experience. With these parameters, K-Means clustering and Density-Based Spatial Clustering of Applications with Noise yielded the Silhouette Coefficients of 0.532 and 0.444 respectively. Such segmentation helps the marketing and sales team of a firm to plan and deliver personalized offers, promotions, rewards, and schemes to attract more customers and thereby increase revenues. The experimental analysis suggested that both methods are very effective in customer segmentation. The Silhouette Coefficient indicates the superiority of K-Means, but, Density-Based Spatial Clustering of Applications with Noise is more explanatory. The work also highlighted the benefits, differences, and suitability of both techniques.