A Kameswara Rao

https://doi.org/10.64823/ijter.2604016

The rapid expansion of digital advertising has created a lucrative target for fraudulent actors who exploit the pay-per-click model by generating illegitimate clicks through automated bots, click farms, and malicious scripts. These fraudulent activities distort campaign analytics, exhaust advertiser budgets, and progressively erode trust in online advertising platforms. Rule-based detection systems — which rely on fixed thresholds and manually defined heuristics — are easily circumvented by adversaries who deliberately engineer their traffic to remain below detection limits while still causing meaningful financial damage. This paper presents a modular, data-driven system that addresses this challenge by combining a 21-feature behavioural representation with a trained XG-Boost gradient boosting classifier to detect fraudulent clicks in real time. The system handles the severe class imbalance inherent to fraud datasets through cost-sensitive learning and threshold optimization guided by the precision-recall curve. Transparency is embedded at the core: SHAP-based explainability generates per-prediction feature-level rationales that are surfaced directly on the advertiser dashboard, converting opaque model decisions into actionable human-readable insights. The complete solution is implemented as a full-stack web application with a React frontend, Node.js/Express backend, Python Flask ML microservice, and MongoDB data layer. Evaluated on the large-scale TalkingData AdTracking benchmark, the deployed XGBoost model achieves an AUC of 0.9549 and a recall of 0.7673, while the hybrid LightGBM–XGBoost ensemble reaches 0.9815 AUC, demonstrating strong predictive performance and practical deployability

This application represents the Static Thick Client malware scanner that is majorly designed to perform the analysis of the Windows Portable Executable (.exe) before dynamically pushing or executing or installing those applications and giving access to the internal environment of the system. This a light-weight modular level approach which has a multiple-layered architecture, including UI , analysis engine , ML prediction model and 6 different and major parts that helps in analyzing and prediction of malware based on the signature and the patterns existing inside it. Multiple detection techniques layered for matching the Malware rules and easy to identify the suspicious categories across the application. Additionally a machine learning model – “Random Forest” included to combinedly classify the extracted features. This application computes a risk score and divided the elements into 6 major categories and allows the user to view and download the reports and scan finding for the betterment of the application security. A user-friendly GUI helps in Scanning , Visualization. For the Logs and the new scanning data is stored in the PostgreSQL . This provides an safe , efficient and scalable solution for the pre-execution and malware detection.

The rapid expansion of digital advertising has created a lucrative target for fraudulent actors who exploit the pay-per-click model by generating illegitimate clicks through automated bots, click farms, and malicious scripts. These fraudulent activities distort campaign analytics, exhaust advertiser budgets, and progressively erode trust in online advertising platforms. Rule-based detection systems — which rely on fixed thresholds and manually defined heuristics — are easily circumvented by adversaries who deliberately engineer their traffic to remain below detection limits while still causing meaningful financial damage. This paper presents a modular, data-driven system that addresses this challenge by combining a 21-feature behavioural representation with a trained XG-Boost gradient boosting classifier to detect fraudulent clicks in real time. The system handles the severe class imbalance inherent to fraud datasets through cost-sensitive learning and threshold optimization guided by the precision-recall curve. Transparency is embedded at the core: SHAP-based explainability generates per-prediction feature-level rationales that are surfaced directly on the advertiser dashboard, converting opaque model decisions into actionable human-readable insights. The complete solution is implemented as a full-stack web application with a React frontend, Node.js/Express backend, Python Flask ML microservice, and MongoDB data layer. Evaluated on the large-scale TalkingData AdTracking benchmark, the deployed XGBoost model achieves an AUC of 0.9549 and a recall of 0.7673, while the hybrid LightGBM–XGBoost ensemble reaches 0.9815 AUC, demonstrating strong predictive performance and practical deployability.