Citation

Rajagopal, Kayapati and Sungheetha, Akey (2026) Explainable Real-Time Ad Bidding Dashboard: A Transparent Deep Reinforcement Learning Framework for Optimized Digital Advertising Performance. Procedia Computer Science, 282. pp. 2827-2843. ISSN 18770509

Text 30.pdf - Published Version Restricted to Repository staff only Download (458kB)

Abstract

Real-time bidding in programmatic advertising faces critical challenges including opacity in decision-making processes, suboptimal bid price determination, and lack of transparency in campaign performance attribution. Existing solutions demonstrate limitations in explainability with average transparency scores below 42 percent, inadequate real-time performance monitoring capabilities, and insufficient integration of interpretable machine learning frameworks. This research proposes an Explainable RealTime Ad Bidding Dashboard integrating transparent deep reinforcement learning with interpretable visualization mechanisms to optimize bidding strategies while maintaining decision transparency. The proposed methodology combines a novel Attentionbased Bidding Network with Shapley Additive Explanations and real-time performance analytics to achieve bidding accuracy of 94.7 percent, cost-per-acquisition reduction of 38.6 percent, and explainability scores exceeding 89.2 percent across diverse campaign scenarios. The system processes bidding decisions within 12.4 milliseconds while providing interpretable feature importance rankings and counterfactual explanations for stakeholder comprehension. Experimental validation on three real-world advertising datasets demonstrates return on ad spend improvements of 2.87 times compared to baseline approaches, with transparency metrics validated through user comprehension studies achieving 91.4 percent stakeholder satisfaction. The framework addresses critical gaps in algorithmic accountability for digital advertising platforms while maintaining competitive performance metrics essential for commercial deployment.

Item Type:	Article
Uncontrolled Keywords:	Transparent machine learning, Digital marketing analytics
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:	03 Sep 2026 09:10
Last Modified:	03 Sep 2026 09:10
URII:	http://shdl.mmu.edu.my/id/eprint/16651

Downloads

Edit (login required)