naivebayes

🏨 Hotel Review Rating Prediction (NLP with Naive Bayes & Logistic Regression)

This project uses natural language processing to predict hotel review ratings based on user-written review text. The analysis involves cleaning a real-world dataset, visualizing insights, and building two machine learning models: Naive Bayes and Logistic Regression.

🔍 Key Highlights

• Data cleaning & EDA
• Visualizations of ratings and top review cities
• TF-IDF vectorization for better word weighting
• Naive Bayes vs Logistic Regression model comparison
• Evaluation using accuracy, precision, recall, F1-score
• Summary insights and recommendations

📈 Results

• Best performance on frequent rating classes (especially rating 5)
• Logistic Regression outperformed Naive Bayes (F1: 0.49 vs 0.45)
• Rare ratings (6–10) were difficult to predict due to imbalance

📦 Tools Used

• Python (Pandas, Matplotlib, Seaborn)
• Scikit-learn
• Jupyter Notebook (Google Colab)

🚀 Run This Notebook

1. Clone the repo or open the notebook in Google Colab
2. Upload the dataset (hotel reviews.csv)
3. Run cells from top to bottom

📌 Author

Billy Pierre – LinkedIn | GitHub