🎬 Sentiment Analysis on IMDB Movie Reviews

Classify movie reviews as Positive or Negative using Natural Language Processing & Machine Learning

---

📖 Overview

This project builds a binary text classifier that determines whether an IMDB movie review expresses a positive or negative sentiment. It walks through a complete end-to-end NLP pipeline — from raw, messy HTML-laden text all the way to a trained Naive Bayes model — achieving 85.11% accuracy on 10,000 held-out reviews.

---

✨ Features

• 🧹 Text Preprocessing — HTML tag removal, punctuation stripping, lowercasing, stopword removal, and Porter Stemming
• 📊 TF-IDF Vectorization — Converts cleaned text into a numerical feature matrix (top 5,000 terms)
• 🤖 Multinomial Naive Bayes Classifier — Fast, interpretable, and effective for text classification
• 📈 Performance Evaluation — Accuracy score + full classification report (precision, recall, F1)

---

🗂️ Project Structure

Sentiment-Analysis/
│
├── sentimentAnalysis.ipynb   # Main Jupyter / Colab notebook
├── IMDB Dataset.csv          # Dataset (50,000 labeled movie reviews)
└── README.md

---

🔄 Pipeline

Raw Reviews (CSV)
      │
      ▼
┌─────────────────────────────────────┐
│  Text Cleaning  (clean_text)        │
│  • Strip HTML tags  <br/>           │
│  • Remove punctuation & numbers     │
│  • Lowercase                        │
│  • Tokenize                         │
│  • Remove English stopwords         │
│  • Porter Stemming                  │
└─────────────────────────────────────┘
      │
      ▼
 Label Encoding   positive → 1  |  negative → 0
      │
      ▼
 Train / Test Split   80% / 20%
      │
      ▼
 TF-IDF Vectorizer   (max_features = 5,000)
      │
      ▼
 Multinomial Naive Bayes  →  Predictions  →  Evaluation

---

📊 Results

Metric	Negative (0)	Positive (1)	Overall
Precision	0.85	0.85	—
Recall	0.84	0.86	—
F1-Score	0.85	0.85	—
Accuracy	—	—	85.11 %

Evaluated on 10,000 test reviews (4,961 negative · 5,039 positive)

---

🚀 Getting Started

1 — Run in the Cloud (Recommended)

Click the badge below — no local setup needed:

2 — Run Locally

# Clone the repository
git clone https://github.com/harshchill/Sentiment-Analysis.git
cd Sentiment-Analysis

# Install dependencies
pip install pandas nltk scikit-learn

# Launch the notebook
jupyter notebook sentimentAnalysis.ipynb

Note: Place IMDB Dataset.csv in the project root before running. The dataset is available on Kaggle.

---

🛠️ Tech Stack

Library	Purpose
pandas	Data loading & manipulation
nltk	Stopwords corpus & Porter Stemmer
scikit-learn	TF-IDF, train/test split, Naive Bayes, metrics
re	Regex-based HTML & punctuation cleaning

---

📚 Dataset

IMDB Movie Reviews — 50,000 polar (positive / negative) movie reviews sourced from the Internet Movie Database.

Property	Value
Total samples	50,000
Classes	Positive, Negative
Balance	50 % / 50 %
Source	Kaggle IMDB Dataset

---

🤝 Contributing

Contributions are welcome! Feel free to open an issue or submit a pull request.

1. Fork the repository
2. Create your feature branch (git checkout -b feature/improve-model)
3. Commit your changes (git commit -m 'Add logistic regression model')
4. Push to the branch (git push origin feature/improve-model)
5. Open a Pull Request

---

📄 License

This project is open-source and available under the MIT License.

---

Made with ❤️ by harshchill