---

🚀 BTC/USDT Prediction API with Docker and Flask

This project contains a machine learning-based cryptocurrency prediction API using Flask, Docker, and a pre-trained joblib model. The API retrieves recent Bitcoin data using the yfinance library, performs preprocessing, and serves predictions through a simple REST interface.

📋 Table of Contents

• Features
• Tech Stack
• Project Structure
• Installation
• Usage
• Docker
• API Endpoints
• Contributing

✨ Features

• 📊 Bitcoin price prediction using a pre-trained machine learning model.
• 🔧 Preprocessing pipeline including log returns, lags, and binning of features.
• 🛠️ Dockerized application for easy deployment and scalability.
• ⚡ Flask API to serve predictions via HTTP.
• 🧠 Model storage using joblib for fast and efficient prediction serving.

🛠️ Tech Stack

• Flask - Python micro-framework to serve the API.
• Joblib - Efficient serialization of the trained machine learning model.
• yFinance - Data retrieval for cryptocurrency (Bitcoin) prices.
• Docker - Containerization for portability and easy deployment.

📂 Project Structure

.
├── Dockerfile                # Dockerfile to build the API container
├── app.py                    # Flask API implementation
├── RF_BTC_1D.joblib          # Pre-trained machine learning model
├── requirements.txt          # Dependencies for the project
└── README.md                 # Project documentation

🐳 Docker Setup

1. Install Docker: Make sure you have Docker installed on your system. You can download it here.

2. Build the Docker Image:

   docker build -t btc-prediction-api .

3. Run the Docker Container:

   docker run -d -p 5000:5000 btc-prediction-api

4. The API will now be available at http://localhost:5000.

⚙️ Installation

1. Clone the repository:

   git clone https://github.com/kinshuksinghbist/btc-prediction-api.git

2. Navigate to the project directory:

   cd btc-prediction-api

3. Install dependencies:

   pip install -r requirements.txt

4. Run the Flask app locally:

   python app.py

The API will now be running locally at http://localhost:5000.

🚀 Usage

Once the API is up and running, you can use the following endpoint:

GET /

Fetches the latest Bitcoin data and returns a prediction based on the pre-trained model.

Example Request:

curl http://localhost:5000/

Example Response:

{
    "predictions": [1, -1, 1, 1, -1, 1]
}

📈 API Endpoints

HTTP Method	Endpoint	Description
GET	/	Returns predicted Bitcoin price moves

🏗️ Model & Preprocessing

• Data Retrieval: We use the yfinance library to retrieve Bitcoin data for the past 19 days.
• Preprocessing: The model applies several transformations to the data:
  • Log Returns: Calculates the log of the price changes.
  • Lags: Adds lag features to capture temporal dependencies.
  • Binning: Converts continuous features into bins to make predictions easier.

The model is pre-trained and saved in the RF_BTC_1D.joblib file. It's a Random Forest classifier that predicts the future direction of Bitcoin price changes.

🤝 Contributing

We welcome contributions! To get started:

1. Fork the repository.
2. Create a new branch (git checkout -b feature-branch).
3. Make your changes.
4. Push to your branch and submit a pull request.

📄 License

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

---