On the Character of Phishing URLs

Abstract

Phishing attacks resulted in an estimated $3.2 billion dollars worth of stolen property in 2007, and the success rate for phishing attacks is increasing each year [17]. Phishing attacks are becoming harder to detect and more elusive by using short time windows to launch attacks. In order to combat the increasing effectiveness of phishing attacks, we propose that combining statistical analysis of website URLs with machine learning techniques will give a more accurate classification of phishing URLs. Using a two-sample Kolmogorov-Smirnov test along with other features we were able to accurately classify 99.3% of our dataset, with a false positive rate of less than 0.4%. Thus, accuracy of phishing URL classification can be greatly increased through the use of these statistical measures.

Comparison of Classification Algorithms for Detection of Phishing Websites

“…5 4 24,520 138,925 99.78% (Zhao and Hoi, 2013) Classic Perceptron 990,000 10,000 99.49% (Patil and Patil, 2018) Random Forest 26,041 26,041 99.44% (Zhao and Hoi, 2013) Label Efficient Perceptron 990,000 10,000 99.41% (Chen et al, 2014) Logistic Regression 1,945 404 99.40% (Cui et al, 2018) SVM 24,520 138,925 99.39% (Patil and Patil, 2018) Fast Decision Tree Learner REPTree26,041 26,041 99.19% (Zhao and Hoi, 2013) Cost-sensitive Perceptron 990,000 10,000 99.18% (Patil and Patil, 2018) C A R T 5 26,041 26,041 99.15% (Jain and Gupta, 2018b) Random Forest 2,141 1,918 99.09% (Patil and Patil, 2018) J 4 8 6 26,041 26,041 99.03% (Verma and Dyer, 2015) J48 11,271 13,274 99.01% (Verma and Dyer, 2015) P A R T 7 11,271 13,274 98.98% (Verma and Dyer, 2015) Random Forest 11,271 13,274 98.88% (Shirazi et al, 2018) Gradient Boosting 1,000 1,000 98,78% (Cui et al, 2018) Naïve-Bayes 24,520 138,925 98,72% (Cui et al, 2018) C4.5 356,215 2,953,700 98.70% (Patil and Patil, 2018) Alternating Decision Tree 26,041 26,041 98.48% (Shirazi et al, 2018) SVM (Linear) 1,000 1,000 98,46% (Shirazi et al, 2018) CART 1,000 1,000 98,42% (Adebowale et al, 2019) Adaptive Neuro-Fuzzy Inference System 6,843 6,157 98.30% (Vanhoenshoven et al, 2016) Random Forest 1,541,000 759,000 98.26% (Jain and Gupta, 2018b) Logistic Regression 2,141 1,918 98.25% (Patil and Patil, 2018) Random Tree 26,041 26,041 98.18% (Shirazi et al, 2018) k-Nearest Neighbuors 1,000 1,000 98,05% (Vanhoenshoven et al, 2016) Multi Layer Perceptron 1,541,000 759,000 97.97% (Verma and Dyer, 2015) Logistic Regression 11,271 13,274 97.70% (Jain and Gupta, 2018b) Naïve-Bayes 2,141 1,918 97.59% (Vanhoenshoven et al, 2016) k-Nearest Neighbours 1,541,000 759,000 97.54% (Shirazi et al, 2018) SVM (Gaussian) 1,000 1,000 97,42% (Vanhoenshoven et al, 2016) C 5 . 0 8 1,541,000 759,000 97.40%…”

A Survey on Phishing Website Detection Using Hadoop

“…However, the accuracy of the login window must match the features provided. According Rakesh Verma and Keith Dyer, proposed a set of lexical URLs, and also how many letters are in them (Verma & Dyer, 2015). However, if the URL does not have spelling errors, then this feature may not work properly.…”

AI@ntiPhish — Machine Learning Mechanisms for Cyber-Phishing Attack

“…Verma et al [22] used four different sources of datasets for training and analysis, including PhishTank.com, APWG, and the DMOZ Open Directory Project [23]. The authors used the logistic regression [24], J48 decision tree [18], and random forest algorithms [25], and analyzed the accuracy of the models through five crossvalidations. Finally, the authors evaluated more relevant feature combinations and trained the most effective detection models.…”

“…The results revealed that 10 was the optimal number of hidden nodes. In addition, the number of weaker classifiers were set to 5,10,15,20,25,30,35,40,45, and 50 for the bagging algorithm and implemented 10 simulations for each given estimator size. The testing accuracies for each estimator size are shown in Fig.…”