Traceability and Fraud in Digital Advertisement

Abstract

In the dynamic landscape of digital advertising, fraud and the lack of traceability pose significant challenges, leading to financial losses and eroding trust among stakeholders. This thesis explores the application of machine learning (ML) techniques to mitigate fraud and enhance traceability within digital ad transactions, including impressions, clicks, and conversions. Through preprocessing, feature engineering, and deploying advanced ML algorithms like XGBoost, this study meticulously evaluates the efficacy of various models in detecting fraudulent activities, employing metrics such as accuracy, precision, recall, F1 score, and ROC-AUC for comprehensive assessment. This research demonstrates that ML significantly outperforms traditional rule-based systems in identifying complex fraudulent patterns, offering a novel framework for integrating ML-based fraud detection into digital advertising platforms. By improving the detection of ad fraud, this work contributes to creating more secure, transparent, and efficient digital advertising markets, ultimately fostering trust and reducing financial losses for advertisers. This thesis not only highlights the financial and trust implications of ad fraud but also presents a methodological approach for leveraging technology to combat these issues, marking a significant step forward in the ongoing battle against digital advertising fraud.