Examining the Behaviors and Preferences of Online Shopping Customers Using Clustering Techniques

Abstract

In recent years, the e-commerce sector has seen a rise in company competitiveness as e-commerce systems have expanded and been used by a variety of businesses. The e-commerce system is an online platform for selling and promoting products to customers. Customer segmentation is a technique of putting customers into groups based on shared traits. The purpose of customer segmentation is to determine a company's target market's purchasing habits, identify trends among various customer segments, and assess customer loyalty. This information helps the business develop targeted marketing campaigns that increase the number of profitable and devoted customers it serves. It is more challenging to secure a customer base in the age of globalization and digitization since consumers have been given many options for making purchases. While obtaining clients is the key objective of any firm, the results of this study and recommendations will assist the organization in developing a range of research criteria to determine the strategies it uses for marketing. Certain unsupervised machine learning techniques were employed in this study to examine customer data. Clusters develop in unsupervised machine learning methods. This business must concentrate and dedicate all its resources to providing superior customer service to its consumers. In this paper, we have employed the technique of principal component analysis (PCA) for dimensionality reduction and k-means, agglomerative, DBSCAN to determine the customer's segment. The findings of this study, k-means algorithms gave a 0.41 silhouette score when they clustered into 5 clusters, the agglomerative algorithms gave a 0.4 silhouette score when they clustered into 4 clusters, and DBSCAN algorithms gave a 0.44 silhouette score when they clustered into 2 clusters. The best model based on the performance is k-means.