Bitcoin Cryptocurrency Forecasting System

Author: Sophie Meyrick-Vernon

Python Version:
Python 3.11

Project Overview:
This project compares LSTM, XGBoost, and Hybrid LSTM-XGBoost models for short term Bitcoin price forecasting using historical cryptocurrency market data collected from the CoinGecko API.

Included Files:
- btc_lstm.py
- btc_xgboost.py
- btc_hybrid.py
- anova_test.py
- bitcoin_data.csv

Required Libraries:
- tensorflow
- xgboost
- numpy
- pandas
- matplotlib
- scikit-learn
- pycoingecko
- scipy

Run Order:
1. Run btc_lstm.py
2. Run btc_xgboost.py
3. Run btc_hybrid.py
4. Run anova_test.py

Notes:
- Models use historical Bitcoin price data in GBP.
- Performance is evaluated using RMSE, MAE, MAPE, R², and directional accuracy.
- ANOVA testing is included for statistical significance analysis.