Using networks and partial differential equations to forecast bitcoin price movement

Area of Science:

• Computational Finance
• Network Science
• Cryptocurrency Analysis

Background:

• Blockchain technology and bitcoin have gained significant attention over the last decade.
• Bitcoin has exhibited substantial volatility in its daily and long-term valuations.
• Existing methods for bitcoin price forecasting often overlook network dynamics and market sentiment.

Purpose of the Study:

• To propose a novel partial differential equation (PDE) model for forecasting bitcoin price movements.
• To analyze the influence of bitcoin transaction patterns on price dynamics.
• To incorporate market sentiment, using Google Trends, into the forecasting model.

Main Methods:

• Development of a PDE model applied to the bitcoin transaction network.
• Analysis of bitcoin subgraphs (chainlets) to understand transaction pattern influences.
• Integration of the Google Trends index to represent market sentiment.
• Validation of the model's forecasting capability through experimental results.

Main Results:

• The proposed PDE model demonstrates capability in forecasting bitcoin price movements.
• The model effectively captures the impact of transaction patterns on price over time.
• Incorporation of Google Trends enhances the model's accuracy by reflecting market sentiment.

Conclusions:

• The PDE model offers a promising new approach for bitcoin price forecasting.
• This research is the first to apply a PDE model to the bitcoin transaction network for price prediction.
• The findings highlight the importance of network topology and market sentiment in understanding bitcoin price dynamics.