Citation

Huq, Sheikh Majedul and Aktar, Shuly and Kumar, Gyanendra and Gopal, R. and Afroz, Sonia and Ali, Md. Akkas (2025) AI-Driven Consumer Behaviour Prediction: A Machine Learning Based Application. In: 5th International Conference on Emerging Research in Electronics, Computer Science and Technology, ICERECT 2025, 12 September 2025 - 13 September 2025, Mandya.

Text IEEE Xplore Full-Text PDF_11.pdf - Published Version Restricted to Repository staff only Download (2MB)

Abstract

Today, fashion retailers face immense pressure to adapt their marketing strategies to keep pace with changing shopper demands, rapid technological advances, and intense competition in the highly connected digital world. Social media has become a critical channel for brand engagement, allowing merchants unprecedented access to insights into consumer preferences and behaviors. However, contrast this with traditional digital marketing methods, which often do not offer the level of personalization, immediacy, and relevance in context needed by the modern-day consumer. AI in digital marketing has now bridged this gap by permitting fashion companies to automate decision-making using large user databases and generate personalized content strategies. This process focuses on the integration of AI in the field of digital marketing, considering in particular the ability of machine learning models to predict consumer decisions and engagement in the fashion retail sector. Given the real user and product-level data from the fashion marketplaces, this work investigates customer purchase intention, user social interaction, and purchase decision metrics using Random Forest Regressor (RF), Support Vector Regression (SVR), and XGBoost, and the features have been pre-processed through engineering, including integration, missing value handling, categorical encoding, and feature scaling. The performance of the proposed method is evaluated using the MAE, MSE, RMSE, and R² metrics. The proposed ML models were more effective than the baseline models with R² of 99.67%, 30.09%, and 99.94%; MAE of (0.0310, 0.0792, 0.0297); MSE of (0.1673, 0.7708, 0.1389); and RMSE of (0.4090, 0.8780, 0.3727) for RF, SVR, and XGBoost, respectively. ANOVA and the Friedman test provisionally confirmed that the differences in model performance were significant. The results show that XGBoost has the best predictive capacity (R² = 99.94%), outperforming Random Forest and SVR. The study contributes to the understanding of AI as it relates to CME profiling and efficient use of social media targeting.

Item Type:	Conference or Workshop Item (Paper)
Uncontrolled Keywords:	Consumer behavior, fashion retail, digital marketing
Subjects:	H Social Sciences > HF Commerce > HF5001-6182 Business > HF5410-5417.5 Marketing. Distribution of products
Divisions:	Faculty of Information Science and Technology (FIST)
Depositing User:	Ms Rosnani Abd Wahab
Date Deposited:	20 Apr 2026 03:25
Last Modified:	20 Apr 2026 03:25
URII:	http://shdl.mmu.edu.my/id/eprint/15768

Downloads

Edit (login required)