Intelligent Association Classification
Technique for Phishing Website Detection
Mustafa Al-Fayoumi1,
Jaber Alwidian2, and Mohammad Abusaif2
1Computer Science Department, Princess Sumaya
University for Technology, Jordan
2Big Data
Department, Intrasoft Middle East, Jordan
Abstract: Many critical applications need more accuracy and
speed in the decision making process. Data mining scholars developed set of
artificial automated tools to enhance the entire decisions based on type of
application. Phishing is one of the most critical application needs for high
accuracy and speed in decision making when a malicious webpage impersonates as
legitimate webpage to acquire secret information from the user. In this paper,
we proposed a new Association Classification (AC) algorithm as an artificial
automated tool to increase the accuracy level of the classification process
that aims to discover any malicious webpage. An Intelligent Association Classification
(IAC) algorithm developed in this article by employing the Harmonic Mean
measure instead of the support and confidence measure to solve the estimation
problem in these measures and discovering hidden pattern not generated by the
existing AC algorithms. Our algorithm compared with four well-known AC
algorithm in terms of accuracy, F1, Precision, Recall and execution time. The
experiments and the visualization process show that the IAC algorithm
outperformed the others in all cases and emphasize on the importance of the
general and specific rules in the classification process.
Keywords: Data mining, Association Classification
technique, Apriori algorithm, Phishing.
Received January 28, 2019; accepted March 28, 2019
https://doi.org/10.34028/iajit/17/4/7