Title: On the use of graph embedding techniques for clustering user browsing navigational behaviours

Authors: Nail Taşgetiren; Talha Kılıç; Mehmet S. Aktas

Addresses: Research and Development Center, Hepsiburada, Istanbul, Turkey ' Research and Development Center, Fibabanka, Istanbul, Turkey ' Computer Engineering Department, Yıldız Technical University, Istanbul, Turkey

Abstract: E-commerce websites offer a multitude of features to their customers through visually appealing interfaces. A significant rise in interest in e-commerce websites has been seen since the onset of the COVID-19 pandemic. Therefore, there is a growing need to incorporate methodologies that can help understand user navigational behaviour on such platforms. This study presents a business process software architecture, which is designed to understand user behaviour through the clustering of similar patterns. To achieve this, graph-based embedding approaches have been utilised. A prototype implementation is presented to demonstrate the proposed business process's efficiency, and the approach's technical details are discussed. Furthermore, the performance of the prototype implementation is evaluated by analysing the quality metrics for clustering. The results of this study show that the proposed business process is successful in analysing and comprehending user navigational behaviour on e-commerce websites.

Keywords: clustering; Word2Vec; Node2Vec; DeepWalk; user gestures; clickstream data; understanding the user behaviour.

DOI: 10.1504/IJWGS.2024.143176

International Journal of Web and Grid Services, 2024 Vol.20 No.4, pp.482 - 504

Received: 14 Oct 2023
Accepted: 21 Mar 2024
Published online: 05 Dec 2024 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article