Phishing and Benign URLs Dataset

Data Set Overview

Data Set Name

Phishing-Website-Detection

Data Set Purpose

Used to train machine learning models and deep neural networks to predict phishing websites.

Data Set Content

• Contains 5,000 phishing URLs and 5,000 legitimate URLs.
• Extracted features include: Domain, Have_IP, Have_At, URL_Length, URL_Depth, Redirection, https, Tiny_URL, Prefix/Suffix, DNS_Record, Web_Traffic, Domain_Age, Domain_End, iFrame, Mouse_Over, Right_Click, Web_Forwards and Label.

Data Processing

• Preprocessing includes feature extraction and data cleaning.
• Dataset split into 80% training and 20% testing.

Usage

• Two Python scripts: one for data preparation, another for implementing and comparing ML algorithms.

Evaluation

• Six ML algorithms: XGboost, Multilayer Perceptrons, Random Forest, Decision Tree, SVM, AutoEncoder.
• Model performance evaluated via predictions on training and test sets, with confusion matrix and visualizations for accuracy comparison.

Conclusion

• XGboost algorithm performs best on the dataset, achieving the highest accuracy.