Cyberfraud Detection on Social Media Analytics Using Machine Learning Ensemble Algorithms in e-Banking

Social media has become a fertile ground for cyberfraud targeting e-banking users. This paper presents a step-by-step methodology for detecting cyberfraud in social media using machine learning algorithms tailored for e-banking contexts. We describe dataset construction (synthetic and real-world sources), feature engineering (textual, network, temporal, and account features), model selection (logistic regression, random forest, XGBoost, and transformer-based models), evaluation metrics (precision, recall, F1, AUC), and explainability techniques (SHAP, LIME). Experimental results on a 1,000-record sample dataset are reported with graphical representations of exploratory data analysis and model performance. The proposed pipeline demonstrates that ensemble tree models and transformer fine-tuning achieve robust detection with explainable decision drivers. Practical considerations, limitations, and suggestions for deployment in e-banking fraud operations are discussed.