0
- Volume 3, Issue 1 2025
By Hania Marfani1 , Hira Kamal1 , Abdul Samad Hussain1 , Darakhshan Syed2
202410.20547/aibd.253103
Keywords: Keywords: K-Means, mini-batch, clustering, customers relationship management, business intelligence, segmentation
Abstract: This study applies k-means and mini-batch clustering dataset for customer segmentation. Based on the analysis, various clusters were formed and validated. It was found that the results of both clustering techniques are same with an enhancement in processing speed via the use of mini-batch k-means. Customer segmentation can be used for market intelligence to identify interested clients by giving corporate entities in the retail sector pertinent and relevant facts periodically. It can be used as methods for examining customer purchase patterns and sales trends. The clustering has been applied over a dataset extracted from Kaggle. After performing the exploratory data analysis, the customers were divided in to platinum, gold, silver and bronze categories. The data were clustered using both the clustering methods (elbow methods were used to cluster the data). The results were found to be same for both clustering techniques. The customers were segmented into platinum, gold, silver and bronze. Results highlighted that mini-batch is more efficient computationally along with providing similar segmentation results as k-means. The validation of the results was performed using elbow method, Silhouette coefficients score was used for further analysis.