Electricity Price Forecasting

Arbitrage in trading is such a powerful strategy to gain profits by spotting market imbalances.

In electricity market, it is realized by ...take positions in the day-ahead market and then offload them in the balancing market (also known as intraday market)....

So, How profitable this strategy can reach, using machine learning algorithms to spot imbalances and turn it into profit&loss signals?

This is a full notebook, with step by step implementation of this strategy to Spanish Electricity market covering:

Getting and preprocessing historical data

• Loading more than 6 years of hourly prices (day-ahead and intraday).

• Cleaning and formatting data to suite model requirements.

Optimizing and Modeling

• Using Optimized (Hyper-parameter tuned) Facebook Prophet Model.

Evaluation

• Evaluate predictions using Cross-validation Technique and sMAPE metric.

Testing Against live data

• Perform Prediction of last 24h of available spreads and compare it with the realized ones.

Have Fun!