Crypto Forecasting Flask App

A web application for predicting cryptocurrency prices using machine learning models.

Overview

This project provides a Flask-based web application that offers cryptocurrency price forecasting using multiple machine learning models. The application supports predictions for Bitcoin (BTC/USDT) and Ethereum (ETH/USDT) using LightGBM, XGBoost, and Linear models.

Project Structure

└── zabih1-cryptoforecasting_flaskapp/
    ├── README.md                    
    ├── app.py                       
    ├── requirements.txt             
    ├── src/                         # Source code
    │   └── ML/                      # Machine learning components
    │       ├── ML_inference.py      # Inference logic for ML models
    │       ├── __init__.py          # Python package initialization
    │       ├── __pycache__/         # Python cached bytecode
    │       └── artifacts/           # Trained model files and assets
    │           ├── model/           # Trained models
    │           │   ├── btcusdt_1d_lgbm_model.pkl     # LightGBM model for BTC
    │           │   ├── btcusdt_1d_linear_model.pkl   # Linear model for BTC
    │           │   ├── btcusdt_1d_xgboost_model.pkl  # XGBoost model for BTC
    │           │   ├── ethusdt_1d_lgbm_model.pkl     # LightGBM model for ETH
    │           │   ├── ethusdt_1d_linear_model.pkl   # Linear model for ETH
    │           │   └── ethusdt_1d_xgboost_model.pkl  # XGBoost model for ETH
    │           └── scaler/          # Feature scalers
    │               ├── btcusdt_1d_scaler.pkl         # Feature scaler for BTC
    │               └── ethusdt_1d_scaler.pkl         # Feature scaler for ETH
    ├── static/                     
    │   └── style.css              
    └── templates/                   
        └── index.html               

Features

• Price predictions for BTC/USDT and ETH/USDT
• Multiple machine learning models:
  • LightGBM
  • XGBoost
  • Linear Regression
• Daily timeframe forecasting
• Web-based user interface

Installation

1. Clone the repository:

   git clone https://github.com/yourusername/zabih1-cryptoforecasting_flaskapp.git
   cd zabih1-cryptoforecasting_flaskapp
   

2. Create a virtual environment (optional but recommended):

   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate
   

3. Install dependencies:

   pip install -r requirements.txt
   

Usage

1. Start the Flask application:

   python app.py
   

2. Open your web browser and navigate to:

   http://127.0.0.1:5000/
   

3. Use the web interface to select:

   • Cryptocurrency (BTC or ETH)
   • Model type (LightGBM, XGBoost, or Linear)
   • Input parameters (if required)

4. View the price prediction results.

Dependencies

The main dependencies include:

• Flask
• NumPy
• pandas
• scikit-learn
• LightGBM
• XGBoost
• pickle

See requirements.txt for the complete list of dependencies and versions.

Models

The application includes pre-trained models for daily price predictions:

• BTC/USDT Models:

  • LightGBM
  • XGBoost
  • Linear Regression

• ETH/USDT Models:

  • LightGBM
  • XGBoost
  • Linear Regression

Each model uses a corresponding scaler to normalize input features.

Development

To extend or modify this application:

• Add new models by placing them in the src/ML/artifacts/model/ directory
• Update the ML_inference.py file to include new inference logic
• Modify the Flask routes in app.py to support new features
• Enhance the UI by updating the templates/index.html and static/style.css files