| dc.description.abstract | Advancements in Information Technology have given rise to an increasingly intercon nected global landscape, simultaneously elevating the criticality of cybersecurity due to
the growing sophistication of cyber threats. Exploiting vulnerabilities within systems
and networks, cybercriminals pose significant risks to confidentiality, integrity, and
availability cornerstones of modern digital infrastructure. Among the various defense
mechanisms, Host-Based Intrusion Detection Systems (HIDS) have emerged as pivotal
tools for detecting and mitigating these evolving threats. Nevertheless, traditional
signature-based detection approaches remain inadequate in addressing contemporary
challenges, including zero-day exploits, ransomware, and Distributed Denial of Service
(DDoS) attacks. This study conducts a systematic review of recent advancements in
HIDS technologies, emphasizing the integration of Machine Learning and Artificial
Intelligence (AI) for anomaly detection and predictive analytics to enable real-time threat
responses. Utilizing PRISMA guidelines, the research synthesizes findings from the
literature to identify key limitations and propose enhancements to HIDS performance.
The analysis reveals that AI-driven models, such as ensemble learning techniques and
adaptive algorithms, significantly enhance detection accuracy, reduce false positive
rates, and improve incident response times. Furthermore, the review underscores the
importance of integrating HIDS with Next-Generation Firewalls (NGFW) to create a
multi-tiered defense framework. NGFWs effectively filter known threats, while HIDS
specialize in identifying complex and sophisticated attack patterns, thereby fostering
resilience against dynamic cyber threats. This paper also outlines future research
directions, including advanced AI integration, cross-network intelligence sharing, and
proactive risk management frameworks, to enhance HIDS capabilities and adapt to the
continuously evolving cyber threat landscape. | en_US |
