## Project background
This repository contains code and results from my student research project (Junior Academy of Sciences of Ukraine, 2024) on **Fractal and Seasonal Analysis of Time Series**.

📄 Full paper: [docs/MAN_Fractal_Analysis.pdf](docs/MAN_Fractal_Analysis.pdf)

# Fractal and Seasonal Time Series Analysis

This repository contains code and results from my Junior Academy of Sciences (МАН) research project on **Fractal Properties and Forecasting of Time Series**.  
The work explores the use of Hurst exponent, Schuster periodogram, ACF/PACF, and SARIMAX/XGBoost models for detecting persistence, cycles, and forecasting.

## Structure
- `docs/` – PDF of the original paper, figures, and abstract
- `src/` – clean Python implementations of methods described in the paper
- `notebooks/` – reproducible experiments and visualizations
- `data/` – sample dataset

## Highlights
- ✅ Hurst exponent via R/S analysis  
- ✅ Schuster periodogram with dominant peak detection  
- ✅ ADF test for stationarity  
- ✅ SARIMAX and XGBoost forecasts  
- ✅ Full reproducibility in Jupyter Notebook

## Quickstart
```bash
pip install -r requirements.txt
jupyter notebook notebooks/analysis.ipynb