Research on Mobile Advertising Click-Through Rate Prediction Algorithm Based on Differential Privacy

Mobile advertising represents a key revenue stream in digital marketing, driven primarily by personalized content. Its effectiveness relies heavily on click-through rate (CTR) prediction models based on user behavior. With growing privacy concerns and increasingly stringent data protection regulations, ensuring user privacy in CTR prediction has become an essential requirement. In response, differential-privacy-based approaches have garnered significant attention. In this study, we propose differentially private mechanisms specifically designed for advanced machine learning models. The approach employs adaptive noise-injection strategies to balance prediction accuracy and privacy effectively. It optimizes the allocation of privacy budgets in CTR estimation while maintaining user anonymity. Experimental results demonstrate that the proposed algorithm achieves prediction accuracy comparable to conventional methods while providing strong privacy guarantees. This framework offers practical solutions that enable mobile advertising platforms to comply with privacy regulations without sacrificing advertising performance.