Feature Engineering and Forecasting via Derivative-free Optimization and Ensemble of Sequence-to-sequence Networks with Applications in Renewable Energy

TL;DR

This paper presents a novel framework combining derivative-free optimization and ensemble sequence-to-sequence networks for improved forecasting and feature engineering in renewable energy, demonstrating superior accuracy especially for long-term predictions.

Contribution

It introduces a new resampling technique, additive resampling, and integrates it with ensemble learning and automated feature selection for renewable energy forecasting.

Findings

The method outperforms existing machine learning techniques in accuracy.

Forecasting accuracy improves with longer horizons.

The framework effectively identifies key environmental features affecting energy output.

Abstract

This study introduces a framework for the forecasting, reconstruction and feature engineering of multivariate processes along with its renewable energy applications. We integrate derivative-free optimization with an ensemble of sequence-to-sequence networks and design a new resampling technique called additive resampling, which, along with Bootstrap aggregating (bagging) resampling, are applied to initialize the ensemble structure. Moreover, we explore the proposed framework performance on three renewable energy sources---wind, solar and ocean wave---and conduct several short- to long-term forecasts showing the superiority of the proposed method compared to numerous machine learning techniques. The findings indicate that the introduced method performs more accurately when the forecasting horizon becomes longer. In addition, we modify the framework for automated feature selection. The model represents a clear interpretation of the selected features. Furthermore, we investigate the effects of different environmental and marine factors on the wind speed and ocean output power, respectively, and report the selected features. Finally, we explore the online forecasting setting and illustrate that the model outperforms alternatives through different measurement errors.