Uniform Manifold Approximation and Projection with Gaussian Mixture Model Based Efficient Clustering Model for Advance Customer Segmentation

Optimization of marketing strategies and enhancing customer engagement have always been a crucial end goal for retail analytics. Data-driven decision making shouldered by Customer Segmentation has played a key role in the field. Traditional clustering methods prove to be insufficient in capturing complex purchasing behavior of customers, more specifically due to any limitations found in dimensionality reduction and clustering accuracy. In this study, we propose an enhanced approach by the integration of Uniform Manifold Approximation and Projection (UMAP) with a fine-tuned Gaussian Mixture Model (GMM) for customer segmentation of a higher efficiency. Conventional methods that rely on Principal Component Analysis (PCA) have been surpassed by our approach using UMAP's superior non-linear feature extraction capabilities to ensure that intricate data structures are preserved. Our results demonstrate a more cohesive segmentation model, offering deeper insights into customer behaviors. This research prepares the way for more adaptive and structure-aware clustering techniques that enhance customer-centric decision making in dynamic retail environments.