Homoglyph Attack Detection Model Using Machine Learning and Hash Function

Phishing is still a major security threat in cyberspace. In phishing, attackers steal critical information from victims by presenting spoofing/fake site that appears to be visual clone of legitimate site. Several Unicode characters are visually identical ASCII characters. This similarity generally known as homoglyphs. Malicious adversaries utilize homoglyphs URLs and DNS domains target organizations. To reduce the risks caused phishing attacks, effective ways detecting websites urgently required. paper proposes homoglyph attack detection model combines hash function machine learning. There two phases approach. The was being trained during development phase. deployment phase involved deploying with Java interface testing outcomes through actual user interaction. results more accurate when URL hashed, any little changes can recognized. detector developed stand-alone software used initial step requesting webpage it enhances browser protects attempts. verify effectiveness, we compared proposed on several criteria existing methods. By using function, features increase overall terms accuracy, integrity, availability. experiment showed detect sites an accuracy 99.8% Random Forest, improves detection.