A MODIFIED ALGORITHM FOR INCREASING THE PERFORMANCE OF MACHINE LEARNING FOR PHISHING ATTACK DETECTION AND CLASSIFICATION
Description
Phishing websites refer to an attack where cyber criminals spoof official websites to lure people into accessing to illegally obtain user identity, password, privacy, and even properties. This attack poses a great threat to inexperienced Internet users and is becoming more and more difficult. Many proposals for detecting phishing websites have shown their effectiveness and the advantage of methods for detecting and classifying Uniform Resource Locators (URLs). Although several approaches have been proposed for detecting and classifying phishing attacks, URL-based machine learning and artificial intelligence approaches provide better performance results, but they all depend on the feature set used. To improve the accuracy of phishing website detection, the article proposes a new decision tree-based model and feature set for Internet of Things (IoT) devices that have limited capabilities and low power consumption. The present study examines how the selection of a feature set from a trainee data set significantly improves the speed and performance of classifying phishing attacks in IoT devices. According to the experimental and comparative results of the implemented classification algorithms, the piecewise linear decision tree algorithm based on the new activation functions provides the best performance with 97.50% accuracy for detecting phishing URLs.
Notes
Files
A-81.pdf
Files
(927.8 kB)
Name | Size | Download all |
---|---|---|
md5:2b548e6b419a64ee62efc5aef500286f
|
927.8 kB | Preview Download |