The adoption of e-banking in has led to significant growth in digital transactions, but it has also increased the vulnerability of the sector to cyber threats. These threats range from phishing and malware attacks to more sophisticated techniques like Advanced Persistent Threats (APTs) and Distributed Denial of Service (DDoS) attacks, which has caused the financial sector to experience substantial losses. As such, traditional cybersecurity measures are proving insufficient to mitigate these risks. This study proposes a cybercrime risk classification and prediction model based on machine learning algorithms, including Decision Tree and Logistic Regression, to enhance cybersecurity risk assessment in e-banking services, to enhance e-banking security classification. Meanwhile A structured questionnaire was designed and administered for data collection, where data preprocessing techniques such as missing values imputation, data normalization, and feature selection to prepare the data for machine learning models was carried out. The dataset was divided into 80% training and 20% testing, for model training and evaluation. Performance metrics, including accuracy, precision, recall, and F1 score, were calculated. The Decision Tree classifier achieved an accuracy of 66.3%, with precision, recall, and F1 scores of 66.8%, 66.2%, and 66.0%, respectively. In comparison, the Logistic Regression classifier obtained a lower accuracy of 56.7%, with corresponding precision, recall, and F1 scores of 56.4%, 56.2%, and 55.9%. The results indicate that the Decision Tree model provides a more effective classification of cybercrime risk levels, offering, proactive approach to identifying and mitigating emerging threats in digital banking sector.

Cybersecurity E-banking Cybercrime Cyber-Attack Machine Learning Security