📊 Sentiment Analysis Web App

An interactive AI-powered web application built with Streamlit that predicts the sentiment of text as positive or negative.
This app supports both Machine Learning and Deep Learning models for sentiment analysis.

---

🚀 Live Demo

🔗 Live App on Streamlit

🚀 Video Demo

Sentiment-Analysis.webm

---

📌 Features

• Predict Sentiment: Detects whether text is positive or negative.
• Dual Model Support: Choose between:
  1. Machine Learning model (.pkl)
  2. Deep Learning model (.keras)
• Interactive UI: Simple input form for entering text.
• Confidence Score: Displays prediction probability for better insights.
• User-friendly Interface: Built using Streamlit with sidebar options.
• Portfolio Links: About Me section included in sidebar.

---

🔍 Usage

1. Open the app in your browser.
2. Select a model from the sidebar:
   • Machine Learning Model (.pkl)
   • Deep Learning Model (.keras)
3. Enter your text in the input area.
4. Click Predict.
5. The app will display:
   • Sentiment: 😊 Positive / ☹️ Negative
   • Confidence Score (0–1)

---

📸 Screenshots

🏠 Home Page

😊 Positive Sentiment Prediction

☹️ Negative Sentiment Prediction

---

⚙️ Tech Stack

• Python 3.9+
• Streamlit (Web App)
• NumPy & Pandas (Data Processing)
• TensorFlow / Keras (Deep Learning Model)
• Scikit-learn & Joblib (Machine Learning Model & Serialization)

---

📄 How It Works

• The text is preprocessed using a tokenizer.
• Converted to sequences and padded to match model input.
• The selected model predicts the sentiment score.
• Score > 0.5 → Positive sentiment; Score ≤ 0.5 → Negative sentiment.

---

👨‍💻 Author

Mirza Yasir Abdullah Baig

• 💼 LinkedIn
• 🌐 Kaggle
• 💻 GitHub

---

❤️ Acknowledgements

• TensorFlow/Keras
• Scikit-learn
• Streamlit Documentation

---

⚠️ Disclaimer

This project is for educational purposes only.
It is not intended for commercial use but demonstrates text-based sentiment analysis with AI.

---

📄 License

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

---