A Predictive Model for Forecasting Locational Marginal Price in Electricity Markets

This praxis develops a predictive model for forecasting Locational Marginal Prices (LMP), considering their time-varying and locational characteristics to deliver valuable price insights for U.S. electricity market participants, including electric utilities and power producers. These insights can aid in formulating trading strategies to manage price fluctuations and risks within LMP-based markets. Statistical and machine learning predictive models were implemented and tested using publicly available electricity market data to forecast LMP across different timeframes, pricing conditions, and locations within the Pennsylvania-New Jersey-Maryland Interconnection (PJM), which operates the world’s largest electricity market. Feature engineering was applied to capture the spatiotemporal nature of LMPs, and a Recursive Feature Elimination (RFE) process was used to rank features, resulting in a simpler and more efficient model. Performance metrics were assessed for all predictive models across various locations, pricing scenarios, and forecasting horizons. This approach enables selecting the most suitable model for specific market conditions, locations, and timeframes for improved decision-making in electricity markets.

Author

• Viray, Zigfred Nino C

Language

• English

Keyword

• Predictive Model
• Recursive Feature Elimination
• Electricity Market
• Machine Learning
• Forecasting
• Locational Marginal Price

Date created

• 2025

Type of Work

• Dissertation

License

• All rights reserved

Rights statement

• In Copyright

GW Unit

• Engineering Management

Degree

• D.Engr.

Advisor

• Biemer, Steven

Committee Member(s)

• Bierce, Jonathan
• Blackford, Joseph P
• Biemer, Steven

Persistent URL

•  https://scholarspace.library.gwu.edu/etd/pv63g124n