📜 Sentiment Analysis using LSTM (Kaggle Dataset)

This repository contains a Sentiment Analysis Model using LSTM (Long Short-Term Memory) on the Kaggle Sentiment Analysis Dataset.
The model classifies reviews as Positive, Negative, or optionally Neutral.

The project includes:

• Training an LSTM model on Kaggle sentiment dataset
• Saving the trained model and tokenizer
• Running predictions on single reviews or batch CSVs
• Support for Google Colab and local execution

---

🔹 Features

• Deep Learning-based LSTM model
• Tokenization and padding for text preprocessing
• Single review prediction
• Colab-ready workflow
• Optional 3-class sentiment support (Positive / Negative / Neutral)

---

📁 Repository Structure

├── Dataset
|    ├── test.csv
|    └── train.csv
├── sentiment_analysis.ipynb  # Colab notebook to train the model
├── tw.keras                  # Trained LSTM model (after training)
├── requirements.txt          # Python dependencies
└── README.md

---

🗂 Dataset

• Dataset: Kaggle Sentiment Analysis Dataset
• CSV columns include:
  • text → The review text
  • label → Sentiment label
    • 0 → Negative
    • 1 → Positive
    • Optional 2 → Neutral (if extended)
• Place the CSV file in the project folder or upload to Colab

---

🛠 Installation

1. Local Setup

git clone https://github.com/yourusername/Sentiment-Analysis.git
cd Sentiment-Analysis
python -m venv venv
source venv/bin/activate   # Windows: venv\Scripts\activate
pip install -r requirements.txt

---

Google Colab Setup

!pip install tensorflow scikit-learn pickle5 pandas

---

🚀 Training the Model

• Load train.csv into the notebook.
• Preprocess text: tokenization and padding
• Convert labels to integers (0,1,2 for neutral if used)
• Train LSTM model
• Save tw.keras
• Test a sample prediction in the notebook

---

📜 License

This project is licensed under the MIT License - see the LICENSE file for details.

---

🤝 Contributing

Feel free to contribute by submitting a pull request or reporting issues!

---

📩 Contact

Ram Narayan Gupta
📧 [email protected]
🌐 LinkedIn Profile