文章基本信息

标题：IMPROVEMENT OF CLASSIFICATION FEATURES TO INCREASE PHISHING TWEETS DETECTION ACCURACY
本地全文：下载
作者：SEOW WOOI LIEW ; NOR FAZLIDA MOHD SANI ; MOHD. TAUFIK ABDULLAH 等
期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2018
卷号：96
期号：10
出版社：Journal of Theoretical and Applied
摘要：Phishing can be defined as a form of social engineering crime that uses to deceive victims by directing them to the fraudulent websites that appear legitimate which will then collect their personal and sensitive information. Phishing attacks use to target email users traditionally but now, target to Online Social Networks (OSN)s typically Twitter. Therefore, a research study of improving classification features for machine learning technique to classify a dataset collected from Twitter is required. In this study, 3 supervised machine learning techniques - Support Vector Machine (SVM), K-Nearest Neighbor (KNN) and Random Forest (RF) and classification features were used to test on a dataset collected from Twitter. The result of our experiment showed that with only 11 selected features, we managed to yield 94.75% classification accuracy higher than 94.56% achieved by other researchers who made use of more than 11 features for the same dataset. From the experiment, we also found that RF remained the best machine learning technique compared to SVM and KNN.
关键词：Phishing; Online Social Networks (OSN)s; Twitter; Classification Features; Machine Learning Techniques