Skip to content

Latest commit

 

History

History
106 lines (80 loc) · 2.41 KB

README.md

File metadata and controls

106 lines (80 loc) · 2.41 KB
Raw

🎉 Sentiment Classifier App 💬

Welcome to the Sentiment Classifier App! This app takes in a review and predicts whether it's positive or negative using machine learning. 🚀

📂 Dataset

We used a dataset of 84,166 reviews to train our model. The reviews are classified as positive or negative based on their content and score.

🔍 Features

  • 🧠 Trained on Logistic Regression – Achieving 91.22% accuracy!
  • 🏋️ TF-IDF Vectorizer – Converts text to numerical features.
  • ⚡ Real-time Sentiment Prediction – Input your review and get instant feedback.

🛠️ How It Works

1. Train the Model

python train_model.py
  • Loads the dataset, preprocesses reviews, trains the classifier, and saves the model.

2. Use the Model

python app.py
  • Launch the interactive app where you can input reviews and receive predictions.

🚀 Getting Started

🧩 Prerequisites

  • Python 3.x
  • Pandas, Scikit-Learn, Joblib

🔧 Installation

  1. Clone the repo: 
    git clone https://github.com/your-username/sentiment-classifier.git
  2. Install dependencies: 
    pip install -r requirements.txt
  3. Train the model: 
    python train_model.py
  4. Start the app: 
    python app.py

⚙️ File Structure

.
├── train_model.py      # Model training script
├── predict_model.py    # Prediction module
├── app.py              # User interface script
├── reviews.csv         # Dataset file (84,166 reviews)
├── sentiment_model.pkl # Trained model
├── vectorizer.pkl      # TF-IDF vectorizer
└── README.md           # This file 😎

🤔 How to Use

  1. Run the app with: 
    python app.py
  2. Enter a review, and the app will predict whether it's positive or negative!
Enter a review: "I love this app, it’s amazing!"
The review is: positive 👍

🧠 Model Performance

Achieved 91.22% accuracy on the dataset with Logistic Regression!

🌟 Contributing

Feel free to fork this repo and contribute! We welcome improvements and new features.

📧 Contact

For any inquiries, reach out at:

Give this repo a ⭐ if you find it helpful! 😊