Comparative Study of Unsupervised Learning Algorithms for Customer Segmentation

This study explores unsupervised learning methods for consumer segmentation—mean shift, hierarchical, and k-means clustering—in the context of vital business-customer interactions. Focused on addressing the escalating demand for extensive data from online platforms, particularly amid growing informational purchases and the research utilizes a Dataset for the exploration. Emphasizing k-means clustering's significance, especially in identifying income-based consumer groups, the study validates its superiority through evaluation metrics (Silhouette Score, Calinski Harabasz Score, and Davies Bouldin Score). Noteworthy insights into stability measurement and the selection of evaluation criteria are provided for future applications, offering valuable guidance to businesses enhancing consumer segmentation in dynamic online environments.