Exponentially-Weighted-Moving-Linear-Regression (EWMLR)

This repository implements the novel Exponentially Weighted Moving Linear Regression (EWMLR) model, which extends the traditional Exponentially Weighted Moving Average (EWMA) by integrating linear regression for improved time series analysis.

EWMLR emphasizes recent observations while capturing broader linear trends, making it highly effective for volatile time series data and real-time decision-making.

Key Features

• Enhanced Forecasting Accuracy: Combines exponential weighting and linear regression for better modeling of dynamic patterns.
• Real-Time Applications: Suitable for financial analysis, demand forecasting, and other dynamic environments.
• Demonstrated Use Case: Evaluated on Bitcoin stock price data, showcasing its ability to outperform traditional EWMA models.

Technologies Used

• Python: Core programming language for implementation.
• Libraries: NumPy, Pandas, Matplotlib, and Seaborn for data processing and visualization.

Future Work

• Extending the model to handle multivariate time series data.
• Exploring applications in other domains like supply chain optimization and IoT.