https://github.com/shervinnd/imdb-sentiment-analysis-lstm

🔍 Dive into sentiment analysis with this LSTM model on the IMDB dataset! Classify movie reviews as positive or negative with high accuracy. 📊 Features ROC curves, custom predictions, and TensorFlow-powered training. Explore code, visualizations, and results.🚀
https://github.com/shervinnd/imdb-sentiment-analysis-lstm

binary-classification data-science deep-learning imdb-dataset jupyter-notebook lstm machine-learning natural-language-processing neural-networks nlp python roc-curve sentiment-analysis tensorflow text-classification

Last synced: 20 days ago
JSON representation

🔍 Dive into sentiment analysis with this LSTM model on the IMDB dataset! Classify movie reviews as positive or negative with high accuracy. 📊 Features ROC curves, custom predictions, and TensorFlow-powered training. Explore code, visualizations, and results.🚀

• Host: GitHub
• URL: https://github.com/shervinnd/imdb-sentiment-analysis-lstm
• Owner: shervinnd
• License: mit
• Created: 2025-09-08T08:31:13.000Z (26 days ago)
• Default Branch: main
• Last Pushed: 2025-09-08T08:41:56.000Z (26 days ago)
• Last Synced: 2025-09-08T10:29:56.797Z (26 days ago)
• Topics: binary-classification, data-science, deep-learning, imdb-dataset, jupyter-notebook, lstm, machine-learning, natural-language-processing, neural-networks, nlp, python, roc-curve, sentiment-analysis, tensorflow, text-classification
• Language: Jupyter Notebook
• Homepage:
• Size: 86.9 KB
• Stars: 2
• Watchers: 0
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
# IMDB Sentiment Analysis with LSTM 🚀

![IMDB Sentiment

Banner](https://img.shields.io/badge/TensorFlow-2.x-orange?style=flat&logo=tensorflow)

![Python](https://img.shields.io/badge/Python-3.8%2B-blue?style=flat&logo=python)

![License](https://img.shields.io/badge/License-MIT-green?style=flat)

🔍**A deep learning project for binary sentiment classification on IMDB

movie reviews.**\

This Jupyter Notebook builds an LSTM-based neural network to analyze

sentiments (positive/negative) from the IMDB dataset. It includes data

preprocessing, model training, evaluation with ROC curves, and custom

review predictions.📈

## 🌟 Features

-   **Dataset Handling**: Loads and preprocesses the IMDB dataset (25K

    training + 25K test reviews). 📚

-   **Model Architecture**: Embedding + LSTM + Dense layers with dropout

    for robust performance. 🧠

-   **Training & Evaluation**: Trains for 5 epochs, evaluates accuracy

    (\~84%), and plots ROC curves (AUC \~0.93). 📊

-   **Custom Predictions**: Test the model on your own movie reviews!

    ✅/❌

-   **Visualizations**: Accuracy plots and ROC curves using Matplotlib.

    🎨

## 🛠️ Installation

1.  Clone the repo:

    ``` bash

    git clone https://github.com/shervinnd/IMDB-Sentiment-Analysis-LSTM.git

    cd IMDB-Sentiment-Analysis-LSTM

    ```

2.  Install dependencies (use a virtual environment recommended):

    ``` bash

    pip install tensorflow numpy matplotlib scikit-learn

    ```

3.  Open the Jupyter Notebook:

    ``` bash

    jupyter notebook IMBD.ipynb

    ```

## 🚀 Usage

1.  **Run the Notebook**: Execute cells sequentially to load data,

    build/train the model, and evaluate.

2.  **Custom Review Testing**: Use the `encode_text` function for new

    reviews. Example:

    ``` python

    sample_text = "This movie was amazing!"

    encoded = encode_text(sample_text)

    prediction = model.predict(encoded)

    print("Positive ✅" if prediction[0][0] > 0.5 else "Negative ❌")

    ```

3.  **Visualize Results**: Check the ROC curve plot for performance

    insights. 📉

## 🧩 Model Architecture

``` plaintext

- Embedding Layer: 10,000 vocab size, 32 dimensions, input length 200.

- LSTM Layer: 64 units.

- Dense Layer: 64 units (ReLU) + Dropout (0.5).

- Output: Sigmoid for binary classification.

```

Compiled with Adam optimizer and binary cross-entropy loss. Achieves

\~84% test accuracy after 5 epochs! 💪

## 📊 Results

-   **Test Accuracy**: 84.26% ✅

-   **ROC AUC**: \~0.93 (visualized in the notebook).\

    Example Predictions:

-   "The movie was fantastic and I really enjoyed it" → Positive ✅

-   "The movie was terrible and I absolutely hated it" → Negative ❌

Training History (sample):

  Epoch   Train Acc   Val Acc

  ------- ----------- ---------

  1       55.78%      82.80%

  5       95.31%      84.38%

## 🤝 Contributing

Pull requests welcome! For major changes, open an issue first. 😊

1.  Fork the repo.

2.  Create a feature branch (`git checkout -b feature`).

3.  Commit changes (`git commit -m`).

4.  Push to the branch (`git push origin feature`).

5.  Open a Pull Request.

## 📄 License

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

for details.