---

Malicious & Spam Links Detector by Delhi Police

This project is a web application that detects whether a given URL is benign or malicious. It uses a machine learning model to classify URLs into these two categories. The project is designed with a user-friendly interface and integrates the branding of the Delhi Police.

Features

• URL Classification: Classifies URLs as either benign or malicious.
• User Interface: A simple web interface where users can input a URL and get the classification result.
• Styled Results Page: The result page features a background with the Delhi Police logo and a styled prediction box.
• Interactive UI: Includes a "Go Back" button on the result page for easy navigation.

Project Structure

.
├── app.py                 # The Flask application code
├── malicious_phish.csv    # The dataset used for training the model (Couldn't upload due to size restrictions)
├── model.pkl              # The trained machine learning model (Couldn't upload due to size restrictions)
├── templates
│   ├── index.html         # The home page where users can input a URL
│   └── result.html        # The results page showing the prediction
├── static
│   ├── style.css          # CSS file for additional styling
│   └── dpimg.jpg          # Delhi Police logo image
└── README.md              # Project documentation

Installation

1. Clone the repository:

   git clone https://github.com/your-username/malicious-link-detector.git
   cd malicious-link-detector

2. Create a virtual environment:

   python -m venv venv

3. Activate the virtual environment:

   • On Windows:

     venv\Scripts\activate

   • On macOS/Linux:

     source venv/bin/activate

4. Install the required packages:

   pip install -r requirements.txt

5. Run the application:

   python app.py

6. Open the browser:

   • Go to http://127.0.0.1:5000 to access the application.

Dataset

The dataset used is malicious_phish.csv, which contains more than 600,000 URLs labeled as benign or malicious. The dataset is used to train the machine learning model.

Machine Learning Model

• Model: RandomForestClassifier
• Features: The model uses a TF-IDF vectorizer to convert URLs into numerical features before classification.
• Training: The model is trained on the dataset, and the trained model is saved as model.pkl.

Usage

1. Input a URL: On the home page (index.html), enter the URL you want to classify.
2. View Results: The result page (result.html) will display the prediction (benign or malicious), along with a "Go Back" button for easy navigation.

Styling

• Background: The result page has a black background with a transparent image of the Delhi Police logo behind the prediction box.
• Prediction Box: A centered box that shows the prediction with a "Go Back" button in red for returning to the home page.

Issues and Troubleshooting

• Always Predicting 'benign' or 'unknown': This issue can occur due to model overfitting or imbalance in the dataset. Ensure the dataset is balanced and the model is trained properly.
• TemplateNotFound Error: Ensure that the index.html and result.html templates are correctly placed in the templates directory.