Evaluating Performance of Intrusion Detection System using Support Vector Machines: Review

Abstract

In the recent work environment, any firm or organization can't guarantee the achievement of its strategic goals, that is, its success without utilizing IT well, so the IT became the object of main attention, and the reliance on the IT gets more intensified in all business fields. Therefore, each firm and public institute take enormous efforts for setting their internal audit systems appropriate for the recent business environment. Especially as the Public Audit Act began to take in effect from 2010, its raised the necessity to develop a new internal audit administration meeting a more advanced internal audit body's role and its status as well as the changes in the administration different from the past administration. Reflecting the past, the existing internal audit activities have done around the post-exposure approaching to the known risks or a kind of fragmentary checkup around samples. Besides, the existing internal audits have focused on pointing out the organization's problems and controlling them. So the organization being audited did not fully accept the internal audit results and there were also some aspects not meeting the internal customer's needs. Therefore this study aimed to propose an improved model of audit information system supporting the recent audit trends (enhancing the efficiency of work process and consulting) in order to draw out a more strategic improvement plan after analyzing the status quo of public firms' internal audits and their problems through the review of existing studies and literature about the theme.

A Survey of Artificial Intelligence in Cyber Security

“…By using cyber-espionage, phishing emails, and other techniques, the attacker takes advantage of these vulnerabilities to use malicious codes or applications. Once the initial breach is done, the attacker maintains control over that [16]. On gain a firm footing, the attacker installs a backdoor or downloads malware to the infected system.…”

“…Examples include buffer overflow, load modules, perl, rootkits, ps, sqlattacks, and x terms. d) Probe: A probe is a program or other device that is put at a crucial junction in a network in order to track or gather information about network activity [16]. For instance, nmap, portsweep, Satan, and ipsweep…”

Application of Machine Learning Approaches in Intrusion Detection System

“…The training dataset of KDD Cup-99 comprises approximately 4,900,000 individual link vectors, each with 41 features and labeled as regular or as an attack, with precisely one particular form of attack [36]. The data collection includes 24 attack forms (connections) falling into one of the four main categories: denial of service (DOS), sample/scanning, root user (U2R), and user remote (R2L).…”

Smart Intrusion Detection System Comprised of Machine Learning and Deep Learning

“…Support Vector Machine (SVM) is used, as a method in a study, to evaluate IDS [27]. Among various approaches of IDS, SVM acts as a classifier with false alarm and detection rate as a measure of performance.…”