Title: Smart and adaptive website navigation recommendations based on reinforcement learning
Authors: I-Hsien Ting; Ying-Ling Tang; Kazunori Minetaki
Addresses: Department of Information Management, National University of Kaohsiung, Taiwan ' Department of Information Management, National University of Kaohsiung, Taiwan ' Kindai University, Japan
Abstract: Improving website structures is the main task of a website designer. In recent years, numerous web engineering researchers have investigated navigation recommendation systems. Page recommendation systems are critical for mobile website navigation. Accordingly, we propose a smart and adaptive navigation recommendation system based on reinforcement learning. In this system, user navigation history is used as the input for reinforcement learning model. The model calculates a surf value for each page of the website; this value is used to rank the pages. On the basis of this ranking, the website structure is modified to shorten the user navigation path length. Experiments were conducted to evaluate the performance of the proposed system. The results revealed that user navigation paths could be decreased by up to 50% with training on 12 months of data, indicating that users could more easily find a target web page with the help of the proposed adaptive navigation recommendation system.
Keywords: web usage mining; adaptive website; navigation recommendation; reinforcement learning.
DOI: 10.1504/IJWGS.2024.139763
International Journal of Web and Grid Services, 2024 Vol.20 No.3, pp.253 - 265
Received: 22 Sep 2023
Accepted: 12 Dec 2023
Published online: 05 Jul 2024 *