Cryptocurrency Prediction Model

This repository presents a data science project focused on forecasting future changes in cryptocurrency prices, leveraging machine learning techniques and time series analysis. The project aims to equip investors, traders, and financial institutions with predictive insights to facilitate informed decision-making in the volatile cryptocurrency market.

Google Colab Notebook

Key Features

1. Machine Learning Modeling:

Two main modeling approaches are utilized - a Sequential neural network model for price prediction based on specific features, and a VARMAX statistical model for time series forecasting to estimate potential future values of those features.

Sequential Model Architecture

2. Interactive Visualization:

An interactive chart is implemented to display both the observed and forecasted cryptocurrency prices, allowing users to access date and price by moving a cursor.

Prerequisites

Libraries

• re
• numpy
• pandas
• matplotlib
• bokeh
• sklearn
• keras
• statsmodels

Data

• Cryptocurrency historical data was collected from investing.com to gather a comprehensive dataset of Bitcoin exchange rates over time.
• The file includes various features such as open, high, low, close prices, trading volume, and percentage changes.
• Any historical cryptocurrency data from this site can be used in the model.

Outcome

• The neural network and time series forecasting models successfully predicted future Bitcoin prices trends until today with reasonable accuracy.
• With the model being relatively raw with no external parameters, relying solely on cryptocurrency source data, this level of accuracy is significant.

Getting Started

Obtaining the data:

• You can download data of a variety of cryptocurrencies from investing.com website.
• Bitcoin data can be obtained from the following link: https://www.investing.com/crypto/bitcoin/historical-data

Running from Source Code:

1. Clone this repository to your local machine.

   git clone https://github.com/arachnocid/Cryptocurrency-Prediction-Model.git

2. Navigate to the project directory.

   cd Cryptocurrency-Prediction-Model

3. Install the required dependencies (see requirements.txt)

   pip install -r requirements.txt

4. Run the "Cryptocurrency-Prediction-Model.py" script.

License

This project is licensed under the GPL-3.0 license.

Author

Arachnocid