SMS Spam Detection App

A Machine Learning web application that detects whether an SMS or Email message is Spam or Not Spam using Natural Language Processing (NLP) and TF-IDF Vectorization.
The application is built using Python, Scikit-learn, and Streamlit, allowing users to enter any message and receive instant predictions.

---

Features

• Real-time spam detection
• Clean NLP preprocessing pipeline
• Interactive web interface using Streamlit
• Lightweight and fast prediction system

---

App Interface

---

Model Performance

• Algorithms: Bernoulli Naive Bayes + Support Vector Machine + ExtraTree Elassifier using Voting algo.

• Feature Extraction: TF-IDF Vectorizer

• Problem Type: Binary Classification (Spam / Not Spam)

• Accuracy: 0.9845261121856866

• Precision: 1.0

---

Machine Learning Workflow

Pipeline

Text → Preprocessing → TF-IDF Vectorization → Trained ML Model → Prediction

1. Data Preprocessing

• Convert text to lowercase
• Tokenization using NLTK
• Remove punctuation and special characters
• Remove stopwords
• Apply stemming using SnowballStemmer

2. Feature Engineering

• Used TF-IDF Vectorizer to convert text into numerical vectors

3. Model Training

• Algorithm: Bernoulli Naive Bayes
• Library: Scikit-learn
• Model serialization using Pickle

4. Deployment

• Built an interactive web app using Streamlit

  Streamlit App --> https://sms-spam-classification-app-aiml.streamlit.app/

---

Tech Stack

• Python
• Scikit-learn
• NLTK
• Streamlit
• Pandas
• NumPy
• Pickle

---

Project Structure

SMS-Spam-Detection/
│
├── app.py                  # Streamlit web app
├── model.pkl               # Trained ML model
├── vectorizer.pkl          # TF-IDF vectorizer
├── Sms_spam_detection.ipynb # Model training notebook
├── requirements.txt        # Project dependencies
└── README.md               # Project documentation

---

Installation & Setup

1. Clone Repository

git clone https://github.com/Vinay-Rai/SMS-spam-classification-APP.git
cd SMS-spam-classification-APP

2. Install Dependencies

pip install -r requirements.txt

3. Run Application

streamlit run app.py

Application runs at:

http://localhost:8501

---

Example Messages

Spam Messages

• "Congratulations! You have won ₹10,000. Claim now."
• "Your bank account will be blocked within 24 hours! Verify your details immediately at www.secure-update-login.com"

Not Spam Messages

• "Are you coming to class today?"
• "Let's meet at 6 pm."

---

Deployment

The application is deployed on Streamlit Community Cloud.

---

Future Improvements

• Improve model accuracy using advanced NLP models (Word2Vec, BERT)
• Add prediction confidence score
• Support bulk message classification
• Enhance UI design
• Email spam detection support

---

Author

Vinay

• Machine Learning Enthusiast
• Interested in NLP and Deep Learning

Github -- > https://github.com/Vinay-Rai

Linkedin--> www.linkedin.com/in/vinay-rai-24vr

---

Support

If you like this project, please give it a ⭐ on GitHub and share it with others!