Bitcoin Price Prediction

Project Overview

This project aims to predict the price of Bitcoin using the following machine learning models:

• Random Forest
• Support Vector Machine
• XGBoost
• ARIMA
• Prophet

The dataset used in this project is the Kaggle dataset Bitcoin Price Prediction (LightWeight CSV)

Table of Contents

• Folder Structure
• Results
• Building
• License

Folder Structure

Bitcoin-Price-Prediction/
│
├── data/
│   ├── case_study_1/
│   │   ├── processed/
│   │   └── raw/
│
├── src/
│   ├── case_study_1/
│   │   ├── model/
│   │   │   ├── arima_functions.py
│   │   │   ├── cross_validation.py
│   │   │   ├── model_evaluation.py
│   │   │   ├── prophet_functions.py
│   │   │   └── walk_forward_validation.py
│   │   ├── preprocessing/
│   │   │   ├── data_cleaning.py
│   │   │   ├── data_preprocessing.py
│   │   │   └── training_models.py
│   │   └── train_evaluate.py
│   │   └── visualization.py
│
├── graphs/
│   ├── case_study_1/
│   │   ├── results/
│   │   │   ├── train_test_split.png
│   │   │   ├── walk_forward_val.png
|
└── config.yaml
└── .gitignore
└── .requirements.txt

Results

The performance of each model is evaluated using multiple metrics and recorded as graphs in graphs/.

Case Study 1

Train, Test Split

Model R² MAPE MAE MSE Prophet 0.368 18.449 413.669 273712.588 XGBoost -2.975 72.076 1278.652 1722022.791 Random Forest -2.990 72.251 1281.341 1728594.680 SVR -3.020 72.415 1285.417 1741651.876 ARIMA -3.761 78.558 1397.157 2062608.181

Time Series Cross Validation

Model	R²	MAPE	MAE	MSE
Prophet	0.812	0.068	97.419	74429.202
SVR	0.539	0.109	164.893	211204.798
Random Forest	0.527	0.117	168.765	206776.694
XGBoost	0.469	0.131	176.450	216565.162
ARIMAX	0.022	0.182	202.285	236345.004

Walk Forward Validation

Model R² MAPE MAE MSE ARIMA 0.979 0.033 64.540 9058.215 Prophet 0.977 0.041 76.843 10018.735 Random Forest 0.510 0.221 343.099 212202.610 XGBoost 0.281 0.248 374.083 311696.006 SVR -2.937 0.719 1273.542 1705706.596

Building

Prerequisites

Before building the project, ensure you have Python installed

Installation

1. Clone the Repository:

   Clone the repository to your local machine using the following command:

   git clone https://github.com/RamezzE/Bitcoin-Price-Prediction.git

2. Navigate to project folder

   cd Bitcoin-Price-Prediction

3. Create & activate virtual environment (Optional but recommended)

   python -m venv venv

   venv\Scripts\activate

4. Install required packages

   pip install -r requirements.txt

5. Run main file to clean & preprocess data then train models and view results

   python src\case_study_1\train_evaluate.py

License

This project is licensed under the MIT License.