IMDB Sentiment Analysis: Multiple NLP Models Comparison

This project is a Natural Language Processing (NLP) pipeline for sentiment analysis on IMDB movie reviews dataset. The goal is to classify reviews as positive or negative and compare the performance of different models.

---

🛠️ Tools & Libraries Used

• Python 3.x
• Jupyter Notebook / Google Colab
• Libraries:
  • pandas → Data manipulation
  • numpy → Numerical operations
  • scikit-learn → Machine learning models & metrics
  • nltk → Text preprocessing (stopwords, lemmatization)
  • matplotlib → Accuracy comparison chart
• Dataset: IMDB Dataset of 50K Movie Reviews

---

🔹 Project Workflow

1. Dataset Loading
   • Load the IMDB dataset (CSV file)
2. Text Preprocessing
   • Lowercase
   • Remove HTML tags
   • Remove non-alphabetic characters
   • Stopwords removal
   • Lemmatization
3. Feature Extraction
   • TF-IDF vectorization (max_features=5000)
4. Train/Test Split
   • 80% training, 20% testing
5. Model Training & Evaluation
   • Logistic Regression
   • Naive Bayes (MultinomialNB)
   • Support Vector Machine (SVM)
   • Random Forest
   • Metrics:
     • Accuracy
     • Classification Report (Precision, Recall, F1-score)
     • Confusion Matrix
6. Accuracy Comparison
   • Bar chart visualization of all models

---

📊 Example Accuracy Results

Model	Accuracy
Logistic Regression	0.88
Naive Bayes	0.85
SVM	0.89
Random Forest	0.86