Shopping Mall Customers Clustering Based on LOF and K-means++

Abstract

Effective customer analysis is vital for business success, and AI-based customer analysis can significantly improve its accuracy, particularly regarding customer grouping and labeling, enabling merchants to generate personalized marketing strategies. Theories on machine learning models for grouping data have been developed since the 1950s, including logistic regression, Support Vector Machine (SVM), decision tree, and random forest; however, computational limitations hindered their practical application. Recent advances in computer technology have led to the development of more accessible machine learning algorithms that generate high-value results. The K-means clustering algorithm is one such model that best fits the customer labeling requirements. As an unsupervised training model, the K-means algorithm clusters customer data into a predetermined number of clusters. In this paper, we apply the K-means algorithm to separately cluster data on male and female clients, while using the K-means++ model to keep initial cluster centers as far apart as possible. We also apply the LOF algorithm to remove any outliers and modify the dataset accordingly.