A Machine Learning Approach to Customers Segmentation Using the K-Nearest Neighbors (knn) Algorithm: an Evaluation of Accuracy and Performance

Student: Isma'il Anchau Suleiman (Project, 2025)
Department of Computer Science and Information Technology
Federal University, Dutsin-Ma, Katsina State

Abstract

In today’s data-driven economy, understanding customer behavior through effective segmentation has become a strategic priority for businesses. This study investigates the application of the KNearest (KNN) algorithm in customer segmentation and evaluates its performance in terms of accuracy and computational efficiency. KNN, a simple yet powerful supervised learning algorithm, is explored for its capacity to group customers based on behavioral and demographic data. The dataset used includes customer attributes such as age, income, spending score and gender, which were preprocessed and standardized prior to modeling. The KNN model was trained and tested using various distance metrics and values of ‘k’ to determine optimal clustering performance. Evaluation metrics such as accuracy, precision, recall, F1-score and computational cost were employed to assess the model. The results indicate that KNN performs reasonable well in segmenting customers, particularly when an optimal ‘k’ value is chosen and data is well-scaled. The study also compares KNN’s performance with other clustering and classification algorithm to provide a broader context for its practical use in marketing analytics.

Keywords
Machine Learning KNN Algorithm Customer Segmentation Data Analysis Model Evaluation Accuracy Performance Classification Predictive Analytics Artificial Intelligence
Filters
Institutions