Spam SMS Detector Web Application

This project is a web application that uses a machine learning model to predict whether an SMS message is spam or ham (not spam). The application is built using Python, Flask (a micro web framework), and scikit-learn for the machine learning model.

Project Structure

    Spam_Mail/
    ├── app.py             # Main Flask application file
    ├── templates/         # Directory for HTML templates
    │   └── index.html     # The main web page
    ├── static/            # Directory for static files (CSS)
    │   └── style.css      # CSS file for styling
    ├── text_preprocessor.joblib   # Saved text preprocessor model
    ├── feature_engineer.joblib    # Saved feature engineering model
    ├── tfidf_vectorizer.joblib    # Saved TF-IDF vectorizer
    ├── spam_classifier.joblib     # Saved trained classification model
    ├── spam.csv             # The original dataset (for training)
    └── README.md            # This file    

Setup and Installation

1. Clone the repository (if applicable):

   git clone https://github.com/Abhay-hack/Spam-Mail.git
   cd Spam_Mail

2. Install required Python libraries:

   pip install -r requirements.txt

   (Note: You might need to create a requirements.txt file if you haven't already. You can generate it using pip freeze > requirements.txt after installing the necessary libraries: pandas, nltk, scikit-learn, imblearn, matplotlib, seaborn, joblib, flask)

3. Download NLTK resources: The script requires some NLTK (Natural Language Toolkit) resources. Run the spam.py script once to download them:

   python spam.py

   This will download punkt, stopwords, and wordnet if they are not already present.

Running the Application

1. Train and Save the Model (if not already done): If you haven't trained and saved the model components (text_preprocessor.joblib, feature_engineer.joblib, tfidf_vectorizer.joblib, spam_classifier.joblib), run the spam.py script:

   python spam.py

   This script will load the data, preprocess it, train a logistic regression model with feature engineering, evaluate it, save the trained components, and then start the Flask development server.

2. Access the Web Application: Once the Flask development server is running (you should see output indicating the server address, usually http://127.0.0.1:5000/), open your web browser and navigate to that address.

Using the Application

1. Enter SMS Message: On the web page, you will find a text area labeled "Enter SMS Message:". Type or paste the SMS message you want to classify into this area.

2. Analyze Message: Click the "Analyze Message" button.

3. View Prediction: The application will process the text using the loaded machine learning model and display the prediction ("Spam" or "Ham") along with a confidence score below the input form.

Model Details

The spam detection model uses the following techniques:

• Text Preprocessing: Lowercasing, URL removal, punctuation removal (except '/'), stop word removal, and lemmatization.
• Feature Engineering: Extraction of text length and presence of capitalized words.
• Text Vectorization: TF-IDF (Term Frequency-Inverse Document Frequency) to convert text into numerical vectors.
• Model: Logistic Regression, trained on a dataset of labeled SMS messages.
• Handling Class Imbalance: SMOTE (Synthetic Minority Over-sampling Technique) is used to address the imbalance between spam and ham messages in the training data.

Libraries Used

• pandas: For data manipulation and analysis.
• nltk: For natural language processing tasks (tokenization, stopwords, lemmatization).
• scikit-learn: For machine learning algorithms (TfidfVectorizer, LogisticRegression, train_test_split, metrics, Pipeline), and base classes for custom transformers.
• imblearn: For handling class imbalance (SMOTE).
• matplotlib and seaborn: For data visualization (confusion matrix).
• joblib: For saving and loading Python objects (the trained model).
• flask: For creating the web application.

Potential Improvements

• More Advanced Feature Engineering: Incorporate more sophisticated features like the presence of specific spam keywords, unusual punctuation patterns, or the ratio of uppercase letters.
• Experiment with Different Models: Try other classification algorithms like Naive Bayes, Support Vector Machines, or ensemble methods.
• User Interface Enhancements: Improve the styling and user experience of the web application.
• Error Handling: Implement more robust error handling for invalid inputs or model loading issues.
• Deployment: Deploy the application to a production server for wider accessibility.

Author

Abhay Gupta