This repository contains Python implementations of essential time series forecasting methods, including:
- ✅ Simple Exponential Smoothing (SES)
- ✅ Double Exponential Smoothing (Holt's Method)
- ✅ Triple Exponential Smoothing (Holt-Winters Method)
- ✅ Moving Averages
The goal of this repository is to demonstrate and compare forecasting techniques using synthetic demand data with trend and seasonality.
- Suitable for data without trend or seasonality.
- Forecasts future values as a constant level, adjusted with a smoothing factor.
- Handles data with a trend.
- Forecast incorporates both level and trend components.
- Handles data with both trend and seasonality.
- Forecasts are adjusted for seasonal patterns (additive and multiplicative).
- Simple method using the mean of past observations.
- Used for short-term forecasts and smoothing series.
- ✅ Self-contained synthetic datasets — no external files required.
- ✅ Clean, well-commented Python code ready for educational and practical use.
- ✅ Automatic forecast extension for future periods.
- ✅ Graphical visualizations for each method.
- ✅ Error analysis: MSE, MAPE, MAD.
- ✅ Statistical tests (Ljung-Box) for autocorrelation checking.
- Python 3.x
- pandas — Data manipulation
- NumPy — Numerical computations
- matplotlib — Plotting and visualization
- scikit-learn — Error metrics (MSE, MAPE, MAD)
- statsmodels — Time series models and statistical tests
- Clone this repository:
git clone https://github.com/RenatoMaynard/time-series-forecasting-methods.git- Install required packages
pip install pandas matplotlib numpy scikit-learn statsmodelsThis project is designed for educational purposes to understand time series forecasting techniques in Python.
This project is for educational purposes only. Errors may exist. Please report issues or contribute improvements via pull requests.
This project is licensed under the MIT License — see the LICENSE file for details.