Enchanced online advertising using ensemble learning

In the era of digital advertising, accurately predicting user engagement is crucial for optimizing ad placements and maximizing revenue. This research introduces an Enhanced Advertising Framework utilizing ensemble learning algorithms to improve ad click predictions. The study followed a structured approach, beginning with data preprocessing, feature engineering, and correlation analysis, ensuring optimal feature selection. The datasets were divided into training and test sets, where models were trained and evaluated using several metrics. The evaluation and testing were conducted using three datasets containing a variety of information, including demographic data, device information, and ad interaction records. XGBoost achieved the highest AUC on AVAZU dataset (0.9333) and DIGIX dataset (0.9743), while AdaBoost excelled on balanced Online Advertising Campaign dataset with an F1-score of 98.52% and AUC of 0.9926. The results highlight the effectiveness of ensemble models in capturing complex patterns in user behavior, leading to more accurate ad targeting.