This repository contains a Python script for phishing detection using machine learning techniques. The script leverages a dataset of URLs labeled as either legitimate or phishing and employs a Decision Tree Classifier to identify potential phishing URLs.
The dataset (malicious_phish.csv) includes URLs along with their corresponding labels indicating whether they are benign or malicious (phishing). The dataset is loaded into a Pandas DataFrame for analysis and model training.
The script extracts various features from the URLs, including domain characteristics, URL length, redirection, presence of IP address, and more. These features are used as inputs to the machine learning model.
- Open the Jupyter notebook (
phishing_detection.ipynb) in an environment that supports execution (e.g., Google Colab). - Execute the notebook cells to load the dataset, extract features, and train the Decision Tree Classifier.
- The script splits the dataset into training and testing sets, trains the model, and evaluates its performance.
- Review the accuracy of the model on the testing set to assess its effectiveness in phishing detection.
malicious_phish.csv: Dataset containing URLs labeled as either benign or phishing.phishing_detection.ipynb: Jupyter notebook containing the script for phishing detection.
The trained model's accuracy on the testing set is computed and displayed. This information helps evaluate the effectiveness of the model in identifying phishing URLs.
The project uses Python, Pandas, scikit-learn, and other libraries for data manipulation, machine learning, and feature extraction. The dataset source is credited within the script.
Feel free to customize this template further based on your project's specific details or any additional information you'd like to include.