Establishing a Dynamic Recommendation System for E-commerce by Integrating Online Reviews, Product Feature Expansion, and Deep Learning

The increasing fragmentation of the consumer journey complicates understanding consumer behavior and tracking digital footprints. Product feature tags should dynamically adapt to consumer preferences, marketing campaigns, and trending internet topics, leveraging crawler technology to address this. H...

Full description

Saved in:
Bibliographic Details
Main Authors: Tsung-Yin Ou, Chun-Hung Chen, Wen-Lung Tsai
Format: Article
Language:English
Published: Taylor & Francis Group 2025-12-01
Series:Applied Artificial Intelligence
Online Access:https://www.tandfonline.com/doi/10.1080/08839514.2025.2463723
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850140123024326656
author Tsung-Yin Ou
Chun-Hung Chen
Wen-Lung Tsai
author_facet Tsung-Yin Ou
Chun-Hung Chen
Wen-Lung Tsai
author_sort Tsung-Yin Ou
collection DOAJ
description The increasing fragmentation of the consumer journey complicates understanding consumer behavior and tracking digital footprints. Product feature tags should dynamically adapt to consumer preferences, marketing campaigns, and trending internet topics, leveraging crawler technology to address this. However, many e-commerce platforms still rely on manual tagging or static product attribute classification, with limited adoption of machine learning approaches. This study proposes a dynamic tagging and recommendation system using deep learning for product image recognition and similarity comparison. By integrating crawler technology, internet trends can serve as dynamic product tags. These tags, combined with consumer behavior data, enable the creation of a recommendation system capable of automatically generating relevant tags. Sales data from 3,132 cartoon products on a Taiwanese e-commerce platform were analyzed. A convolutional neural network was employed to establish a tagging and image recognition model, and it was trialed over 24 weeks. Results showed significant improvements in consumer engagement: average clicks per product increased by 36.1%, views by 22.9%, products added to carts by 32.3%, orders by 28.3%, and payment transactions by 30.4%. Aligning recommendation systems with consumer expectations enhances their ability to identify preferences and drive purchasing behavior.
format Article
id doaj-art-6d9895eb5a834c0fae59d4dff4a404ba
institution OA Journals
issn 0883-9514
1087-6545
language English
publishDate 2025-12-01
publisher Taylor & Francis Group
record_format Article
series Applied Artificial Intelligence
spelling doaj-art-6d9895eb5a834c0fae59d4dff4a404ba2025-08-20T02:29:58ZengTaylor & Francis GroupApplied Artificial Intelligence0883-95141087-65452025-12-0139110.1080/08839514.2025.2463723Establishing a Dynamic Recommendation System for E-commerce by Integrating Online Reviews, Product Feature Expansion, and Deep LearningTsung-Yin Ou0Chun-Hung Chen1Wen-Lung Tsai2Department of Marketing and Distribution Management, National Kaohsiung University of Science and Technology, Kaohsiung, TaiwanCollege of Management, National Kaohsiung University of Science and Technology, Kaohsiung, TaiwanDepartment of Information Management, National Taipei University of Business, Taipei, TaiwanThe increasing fragmentation of the consumer journey complicates understanding consumer behavior and tracking digital footprints. Product feature tags should dynamically adapt to consumer preferences, marketing campaigns, and trending internet topics, leveraging crawler technology to address this. However, many e-commerce platforms still rely on manual tagging or static product attribute classification, with limited adoption of machine learning approaches. This study proposes a dynamic tagging and recommendation system using deep learning for product image recognition and similarity comparison. By integrating crawler technology, internet trends can serve as dynamic product tags. These tags, combined with consumer behavior data, enable the creation of a recommendation system capable of automatically generating relevant tags. Sales data from 3,132 cartoon products on a Taiwanese e-commerce platform were analyzed. A convolutional neural network was employed to establish a tagging and image recognition model, and it was trialed over 24 weeks. Results showed significant improvements in consumer engagement: average clicks per product increased by 36.1%, views by 22.9%, products added to carts by 32.3%, orders by 28.3%, and payment transactions by 30.4%. Aligning recommendation systems with consumer expectations enhances their ability to identify preferences and drive purchasing behavior.https://www.tandfonline.com/doi/10.1080/08839514.2025.2463723
spellingShingle Tsung-Yin Ou
Chun-Hung Chen
Wen-Lung Tsai
Establishing a Dynamic Recommendation System for E-commerce by Integrating Online Reviews, Product Feature Expansion, and Deep Learning
Applied Artificial Intelligence
title Establishing a Dynamic Recommendation System for E-commerce by Integrating Online Reviews, Product Feature Expansion, and Deep Learning
title_full Establishing a Dynamic Recommendation System for E-commerce by Integrating Online Reviews, Product Feature Expansion, and Deep Learning
title_fullStr Establishing a Dynamic Recommendation System for E-commerce by Integrating Online Reviews, Product Feature Expansion, and Deep Learning
title_full_unstemmed Establishing a Dynamic Recommendation System for E-commerce by Integrating Online Reviews, Product Feature Expansion, and Deep Learning
title_short Establishing a Dynamic Recommendation System for E-commerce by Integrating Online Reviews, Product Feature Expansion, and Deep Learning
title_sort establishing a dynamic recommendation system for e commerce by integrating online reviews product feature expansion and deep learning
url https://www.tandfonline.com/doi/10.1080/08839514.2025.2463723
work_keys_str_mv AT tsungyinou establishingadynamicrecommendationsystemforecommercebyintegratingonlinereviewsproductfeatureexpansionanddeeplearning
AT chunhungchen establishingadynamicrecommendationsystemforecommercebyintegratingonlinereviewsproductfeatureexpansionanddeeplearning
AT wenlungtsai establishingadynamicrecommendationsystemforecommercebyintegratingonlinereviewsproductfeatureexpansionanddeeplearning