Design of an AI-Based Dynamic Recommender and Marketing System Using Users’ Sentiment Data from Social Media for E-Commerce

Design of an AI-Based Dynamic Recommender and Marketing System Using Users’ Sentiment Data from Social Media for E-Commerce

Hosein Sojodi1 Hasan Ziafat2

1) Department of Computer Engineering, Electronic Branch, Islamic Azad University, Tehran, Iran
2) Department of Computer, Nat.C., Islamic Azad University, Natanz, Iran

IOI Code : AAHD-ABGHGE
Publication : National Conference on Information Technology, Nanotechnology, Artificial Intelligence and Technological Futures Studies(itnaf.ir)
Abstract :
    In recent years, recommender systems have attracted significant attention as one of the key tools for improving user experience and enhancing the efficiency of e-commerce platforms. However, most of these systems rely primarily on users’ behavioral and transactional data, while the role of user emotions in the decision-making process is often overlooked. The aim of this study is to design and implement an integrated system comprising an intelligent recommender system and a dynamic marketing mechanism based on sentiment analysis of users’ social media data. In this research, textual data from users are first collected from various sources and, after preprocessing, are analyzed using a BERT-based deep learning sentiment analysis model. Then, in order to model relationships between users and products and to capture temporal changes in user behavior, a dynamic graph neural network model is employed. The output of this model is used as the input to the recommender system to provide personalized recommendations aligned with users’ preferences and emotional states. In addition, a dynamic marketing mechanism is designed that leverages users’ emotional information to deliver real-time, personalized marketing messages and offers. The evaluation results demonstrate that the proposed model significantly improves recommendation accuracy, ranking quality, click-through rate, and conversion rate compared to traditional methods. The findings of this study indicate that integrating sentiment analysis with deep learning and dynamic graph-based models can play an important role in developing next-generation recommender systems and lead to a more intelligent, personalized, and user-centric experience in e-commerce platforms.

Keywords : recommender systems sentiment analysis dynamic graph deep learning intelligent marketing