Prediction of solar flares is critical for minimizing the impact of space weather on communication and power systems. This study explores the application of machine learning models—Long Short-Term Memory (LSTM), Random Forest (RF), and XGBoost (XGB)—to forecast solar flares using multivariate time series data derived from SHARP and GOES. The experimental setup systematically varies two key temporal parameters: window size (12, 24, and 48 hours) and prediction lag (12, 24, and 48 hours). To address the strong class imbalance in flare data, cost-sensitive learning is incorporated using a sample-based class weighting strategy. Results demonstrate that the LSTM model achieves the best performance when using a short 12-hour window and a 12-hour prediction lag, reaching 89% accuracy, a True Skill Statistic (TSS) of 0.6949, and an F1-score of 0.6873. These findings emphasize the importance of short-term temporal dependencies and class imbalance mitigation for improving solar flare forecasting performance.

This paper presents a novel approach for forecasting gas consumption in pelletizing processes in the steel industry by integrating AutoML techniques based on neuroevolution. The pursuit of energy efficiency and the reduction of harmful gas emissions is a pressing challenge for industrial sustainability. In this context, we developed and implemented MLP and LSTM neural network models, optimized through neuroevolution strategies, to enhance the accuracy of gas consumption forecasts. The use of AutoML allowed for the automation of model selection and tuning, significantly reducing the need for manual intervention and improving the effectiveness of the predictive models. The results showed that the LSTM model, optimized by neuroevolution, significantly outperformed traditional methods, achieving an RMSE of 0.39, demonstrating not only the superior accuracy of the proposed approach but also its practical relevance for the industry. Additionally, the study highlighted the efficiency of neuroevolution in configuring adaptive network architectures that efficiently respond to the complex dynamics of industrial data. These findings support the adoption of AutoML as a strategic tool for optimizing industrial processes, contributing to the literature on practical applications of advanced machine learning and energy resource management in industrial environments.