Hasan, Mahbub and Rahman, Mohammad Masudur and Sukhi, Rabia Akter and Akter, Fahmida and Bhuiyan, Mohd. Salman Hossain (2024) Predictive Cybersecurity Analytics for Insider Threat Detection in Zero-Trust Enterprise Environments. American Journal of Economics and Business Management, 7 (12). ISSN 2576-5973

Text AJEBM_MAHBUB_Predictive+Cybersecurity+Analytics.pdf Download (1MB)

Abstract

Insider threats are one of the most serious challenges in today's cyber world as enterprise cybersecurity becomes increasingly complex due to the fast-paced adoption of cloud computing, hybrid working models and digital transformation. Internal threats are posed by users with valid access to the network and are thus hard to identify with traditional security controls as they are carried out by authorized users. While Zero-Trust Enterprise Architecture is about constant verification and least-privilege access, Zero Trust also needs to be intelligent to proactively detect suspicious user activity before security incidents happen. This study is aimed at developing a predictive cybersecurity analytics framework for insider threat detection in a Zero-Trust enterprise environment based on machine learning techniques and behavioral analytics. This study leverages the CERT Insider Threat Dataset which includes enterprise email communication records that track user interactions, communication trends, device interactions, attachment activity, and temporal interactions. In the data preprocessing stage, any indicators of activities related to insider threats are identified. In the data preprocessing stage, the meaningful indicators of insider threat activities are extracted. Several machine learning models such as Random Forest, Support Vector Machine (SVM), Decision Tree, XGBoost and isolation forest are used to create predictive models that identify if there is any normal user behavior or if there are any anomalous or potentially malicious activity. To guarantee accurate and solid detection, model performance is assessed by Accuracy, Precision, Recall, F1-score, Receiver Operating Characteristic Area under the Curve (ROC-AUC) and False Positive Rate. The suggested model combines the predictive analysis capabilities of AI with Zero-Trust security principles, allowing for ongoing behavioral monitoring, risk evaluation and threat prediction. The outcomes of these are that it enables better accuracy in detection, fewer false alerts, better security resilience for enterprise and assistance in real-time cybersecurity decision making. In this research, the authors share their experience of making predictive cybersecurity analytics an effective way to bolster ZZEE by leveraging data-driven insights to analyze behaviour and machine learning, and how this enhances the security of modern enterprise networks and reduces the likelihood of insider threats.

Item Type:	Article
Uncontrolled Keywords:	Predictive Cybersecurity Analytics, Insider Threat Detection, Zero-Trust Enterprise Environments, Machine Learning, Behavioral Analytics and CERT inside Threat
Subjects:	H Social Sciences
Depositing User:	admin eprints
Date Deposited:	22 Oct 2026 01:34
Last Modified:	22 Oct 2026 01:34
URI:	http://eprints.umsida.ac.id/id/eprint/17233

Actions (login required)

View Item